PDF Advances in Computer Vision Proceedings of The 2019 Computer Vision Conference CVC Volume 1 Kohei Arai Ebook Full Chapter

Advances in Computer Vision:
Proceedings of the 2019 Computer

Vision Conference (CVC), Volume 1
Kohei Arai
Visit to download the full and correct content document:
https://textbookfull.com/product/advances-in-computer-vision-proceedings-of-the-201
9-computer-vision-conference-cvc-volume-1-kohei-arai/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Proceedings of the Future Technologies Conference FTC

2018 Volume 1 Kohei Arai
https://textbookfull.com/product/proceedings-of-the-future-
technologies-conference-ftc-2018-volume-1-kohei-arai/
Proceedings of the Future Technologies Conference (FTC)

2020, Volume 1 Kohei Arai
Intelligent Computing Proceedings of the 2020 Computing

Conference Volume 1 Kohei Arai
https://textbookfull.com/product/intelligent-computing-
proceedings-of-the-2020-computing-conference-volume-1-kohei-arai/
Applied Computer Vision and Image Processing

Proceedings of ICCET 2020 Volume 1 Brijesh Iyer
https://textbookfull.com/product/applied-computer-vision-and-
image-processing-proceedings-of-iccet-2020-volume-1-brijesh-iyer/
Intelligent Computing: Proceedings of the 2018
Computing Conference, Volume 2 Kohei Arai
Proceedings of the Future Technologies Conference FTC

2018 Volume 2 Kohei Arai


Computer Vision Systems 12th International Conference

ICVS 2019 Thessaloniki Greece September 23 25 2019
Proceedings Dimitrios Tzovaras
https://textbookfull.com/product/computer-vision-systems-12th-
international-conference-icvs-2019-thessaloniki-greece-
september-23-25-2019-proceedings-dimitrios-tzovaras/
Advances in Intelligent Systems and Computing 943
Kohei Arai
Supriya Kapoor Editors
Advances in
Computer
Vision
Proceedings of the 2019 Computer
Vision Conference (CVC), Volume 1
Advances in Intelligent Systems and Computing
Volume 943
Series Editor
Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences,
Warsaw, Poland
Advisory Editors
Nikhil R. Pal, Indian Statistical Institute, Kolkata, India
Rafael Bello Perez, Faculty of Mathematics, Physics and Computing,
Universidad Central de Las Villas, Santa Clara, Cuba
Emilio S. Corchado, University of Salamanca, Salamanca, Spain
Hani Hagras, Electronic Engineering, University of Essex, Colchester, UK
László T. Kóczy, Department of Automation, Széchenyi István University,
Gyor, Hungary
Vladik Kreinovich, Department of Computer Science, University of Texas
at El Paso, El Paso, TX, USA
Chin-Teng Lin, Department of Electrical Engineering, National Chiao
Tung University, Hsinchu, Taiwan
Jie Lu, Faculty of Engineering and Information Technology,
University of Technology Sydney, Sydney, NSW, Australia
Patricia Melin, Graduate Program of Computer Science, Tijuana Institute
of Technology, Tijuana, Mexico
Nadia Nedjah, Department of Electronics Engineering, University of Rio de Janeiro,
Rio de Janeiro, Brazil
Ngoc Thanh Nguyen, Faculty of Computer Science and Management,
Wrocław University of Technology, Wrocław, Poland
Jun Wang, Department of Mechanical and Automation Engineering,
The Chinese University of Hong Kong, Shatin, Hong Kong
The series “Advances in Intelligent Systems and Computing” contains publications
on theory, applications, and design methods of Intelligent Systems and Intelligent
Computing. Virtually all disciplines such as engineering, natural sciences, computer
and information science, ICT, economics, business, e-commerce, environment,
healthcare, life science are covered. The list of topics spans all the areas of modern
intelligent systems and computing such as: computational intelligence, soft comput-
ing including neural networks, fuzzy systems, evolutionary computing and the fusion
of these paradigms, social intelligence, ambient intelligence, computational neuro-
science, artificial life, virtual worlds and society, cognitive science and systems,
Perception and Vision, DNA and immune based systems, self-organizing and
adaptive systems, e-Learning and teaching, human-centered and human-centric
computing, recommender systems, intelligent control, robotics and mechatronics
including human-machine teaming, knowledge-based paradigms, learning para-
digms, machine ethics, intelligent data analysis, knowledge management, intelligent
agents, intelligent decision making and support, intelligent network security, trust
management, interactive entertainment, Web intelligence and multimedia.
The publications within “Advances in Intelligent Systems and Computing” are
primarily proceedings of important conferences, symposia and congresses. They
cover significant recent developments in the field, both of a foundational and
applicable character. An important characteristic feature of the series is the short
publication time and world-wide distribution. This permits a rapid and broad
dissemination of research results.
** Indexing: The books of this series are submitted to ISI Proceedings,

EI-Compendex, DBLP, SCOPUS, Google Scholar and Springerlink **
More information about this series at http://www.springer.com/series/11156

Kohei Arai Supriya Kapoor
•
Editors
Advances in
Computer Vision
Proceedings of the 2019 Computer Vision
Conference (CVC), Volume 1
123
Editors
Kohei Arai Supriya Kapoor
Saga University The Science and Information
Saga, Saga, Japan (SAI) Organization
Bradford, West Yorkshire, UK
ISSN 2194-5357 ISSN 2194-5365 (electronic)

Advances in Intelligent Systems and Computing
ISBN 978-3-030-17794-2 ISBN 978-3-030-17795-9 (eBook)
https://doi.org/10.1007/978-3-030-17795-9
© Springer Nature Switzerland AG 2020
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
It gives us the great pleasure to welcome all the participants of the Computer Vision
Conference (CVC) 2019, organized by The Science and Information
(SAI) Organization, based in the UK. CVC 2019 offers a place for participants to
present and to discuss their innovative recent and ongoing research and their
applications. The prestigious conference was held on 25–26 April 2019 in
Las Vegas, Nevada, USA.
Computer vision is a field of computer science that works on enabling the
computers to identify, see and process information in a similar way that humans do
and provide an appropriate result. Nowadays, computer vision is developing at a
fast pace and has gained enormous attention.
The volume and quality of the technical material submitted to the conference
confirm the rapid expansion of computer vision and CVC’s status as its flagship
conference. We believe the research presented at CVC 2019 will contribute to
strengthen the great success of computer vision technologies in industrial, enter-
tainment, social and everyday applications. The participants of the conference were
from different regions of the world, with the background of either academia or
industry.
The published proceedings has been divided into two volumes, which covered a
wide range of topics in Machine Vision and Learning, Computer Vision
Applications, Image Processing, Data Science, Artificial Intelligence, Motion and
Tracking, 3D Computer Vision, Deep Learning for Vision, etc. These papers are
selected from 371 submitted papers and have received the instruction and help from
many experts, scholars and participants in proceedings preparation. Here, we would
like to give our sincere thanks to those who have paid great efforts and support
during the publication of the proceeding. After rigorous peer review, 118 papers
were published including 7 poster papers.
Many thanks go to the Keynote Speakers for sharing their knowledge and
expertise with us and to all the authors who have spent the time and effort to
contribute significantly to this conference. We are also indebted to the organizing
committee for their great efforts in ensuring the successful implementation of the
v
vi Preface
conference. In particular, we would like to thank the technical committee for their
constructive and enlightening reviews on the manuscripts in the limited timescale.
We hope that all the participants and the interested readers benefit scientifically
from this book and find it stimulating in the process. See you in next SAI
Conference, with the same amplitude, focus and determination.
Regards,
Kohei Arai
Contents
Deep Learning for Detection of Railway Signs and Signals . . . . . . . . . . 1

Georgios Karagiannis, Søren Olsen, and Kim Pedersen
3D Conceptual Design Using Deep Learning . . . . . . . . . . . . . . . . . . . . . 16
Zhangsihao Yang, Haoliang Jiang, and Lan Zou
The Effect of Color Channel Representations on the Transferability
of Convolutional Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Javier Diaz-Cely, Carlos Arce-Lopera, Juan Cardona Mena,
and Lina Quintero
Weakly Supervised Deep Metric Learning for Template Matching . . . . 39
Davit Buniatyan, Sergiy Popovych, Dodam Ih, Thomas Macrina,
Jonathan Zung, and H. Sebastian Seung
Nature Inspired Meta-heuristic Algorithms for Deep Learning:
Recent Progress and Novel Perspective . . . . . . . . . . . . . . . . . . . . . . . . . 59
Haruna Chiroma, Abdulsalam Ya’u Gital, Nadim Rana,
Shafi’i M. Abdulhamid, Amina N. Muhammad, Aishatu Yahaya Umar,
and Adamu I. Abubakar
Transfer Probability Prediction for Traffic Flow with Bike
Sharing Data: A Deep Learning Approach . . . . . . . . . . . . . . . . . . . . . . 71
Wenwen Tu and Hengyi Liu
CanvasGAN: A Simple Baseline for Text to Image Generation
by Incrementally Patching a Canvas . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Amanpreet Singh and Sharan Agrawal
Unsupervised Dimension Reduction for Image Classification
Using Regularized Convolutional Auto-Encoder . . . . . . . . . . . . . . . . . . . 99
Chaoyang Xu, Ling Wu, and Shiping Wang
vii
viii Contents
ISRGAN: Improved Super-Resolution Using Generative

Adversarial Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Vishal Chudasama and Kishor Upla
Deep Learning vs. Traditional Computer Vision . . . . . . . . . . . . . . . . . . 128
Niall O’Mahony, Sean Campbell, Anderson Carvalho,
Suman Harapanahalli, Gustavo Velasco Hernandez, Lenka Krpalkova,
Daniel Riordan, and Joseph Walsh
Self-localization from a 360-Degree Camera Based on the Deep
Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Shintaro Hashimoto and Kosuke Namihira
Deep Cross-Modal Age Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Ali Aminian and Guevara Noubir
Multi-stage Reinforcement Learning for Object Detection . . . . . . . . . . . 178
Jonas König, Simon Malberg, Martin Martens, Sebastian Niehaus,
Artus Krohn-Grimberghe, and Arunselvan Ramaswamy
Road Weather Condition Estimation Using Fixed and Mobile
Based Cameras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Koray Ozcan, Anuj Sharma, Skylar Knickerbocker, Jennifer Merickel,
Neal Hawkins, and Matthew Rizzo
Robust Pedestrian Detection Based on Parallel Channel
Cascade Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Jiaojiao He, Yongping Zhang, and Tuozhong Yao
Novel Scheme for Image Encryption and Decryption Based
on a Hermite-Gaussian Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
Mohammed Alsaedi
MAP Interpolation of an Ising Image Block . . . . . . . . . . . . . . . . . . . . . 237
Matthew G. Reyes, David L. Neuhoff, and Thrasyvoulos N. Pappas
Volumetric Data Exploration with Machine Learning-Aided
Visualization in Neutron Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Yawei Hui and Yaohua Liu
License Plate Character Recognition Using Binarization
and Convolutional Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Sandeep Angara and Melvin Robinson
3D-Holograms in Real Time for Representing Virtual Scenarios . . . . . . 284
Jesús Jaime Moreno Escobar, Oswaldo Morales Matamoros,
Ricardo Tejeida Padilla, and Juan Pablo Francisco Posadas Durán
Contents ix
A Probabilistic Superpixel-Based Method for Road Crack

Network Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
J. Josiah Steckenrider and Tomonari Furukawa
Using Aerial Drone Photography to Construct 3D Models of Real
World Objects in an Effort to Decrease Response Time
and Repair Costs Following Natural Disasters . . . . . . . . . . . . . . . . . . . . 317
Gil Eckert, Steven Cassidy, Nianqi Tian, and Mahmoud E. Shabana
Image Recognition Model over Augmented Reality Based on
Convolutional Neural Networks Through Color-Space Segmentation . . . . 326
Andrés Ovidio Restrepo-Rodríguez, Daniel Esteban Casas-Mateus,
Paulo Alonso Gaona-García, and Carlos Enrique Montenegro-Marín
License Plate Detection and Recognition: An Empirical Study . . . . . . . . 339
Md. J. Rahman, S. S. Beauchemin, and M. A. Bauer
Automatic Object Segmentation Based on GrabCut . . . . . . . . . . . . . . . . 350
Feng Jiang, Yan Pang, ThienNgo N. Lee, and Chao Liu
Vertebral Body Compression Fracture Detection . . . . . . . . . . . . . . . . . . 361
Ahmet İlhan, Şerife Kaba, and Enver Kneebone
PZnet: Efficient 3D ConvNet Inference on Manycore CPUs . . . . . . . . . . 369
Sergiy Popovych, Davit Buniatyan, Aleksandar Zlateski, Kai Li,
and H. Sebastian Seung
Evaluating Focal Stack with Compressive Sensing . . . . . . . . . . . . . . . . . 384
Mohammed Abuhussein and Aaron L. Robinson
SfM Techniques Applied in Bad Lighting and Reflection
Conditions: The Case of a Museum Artwork . . . . . . . . . . . . . . . . . . . . . 394
Laura Inzerillo
Fast Brain Volumetric Segmentation from T1 MRI Scans . . . . . . . . . . . 402
Ananya Anand and Namrata Anand
No-reference Image Denoising Quality Assessment . . . . . . . . . . . . . . . . . 416
Si Lu
Plant Leaf Disease Detection Using Adaptive
Neuro-Fuzzy Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434
Hiteshwari Sabrol and Satish Kumar
Fusion of CNN- and COSFIRE-Based Features with Application
to Gender Recognition from Face Images . . . . . . . . . . . . . . . . . . . . . . . 444
Frans Simanjuntak and George Azzopardi
x Contents
Standardization of the Shape of Ground Control Point (GCP)

and the Methodology for Its Detection in Images for UAV-Based
Mapping Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
Aman Jain, Milind Mahajan, and Radha Saraf
Non-linear-Optimization Using SQP for 3D Deformable Prostate
Model Pose Estimation in Minimally Invasive Surgery . . . . . . . . . . . . . 477
Daniele Amparore, Enrico Checcucci, Marco Gribaudo,
Pietro Piazzolla, Francesco Porpiglia, and Enrico Vezzetti
TLS-Point Clouding-3D Shape Deflection Monitoring . . . . . . . . . . . . . . 497
Gichun Cha, Byungjoon Yu, Sehwan Park, and Seunghee Park
From Videos to URLs: A Multi-Browser Guide to Extract
User’s Behavior with Optical Character Recognition . . . . . . . . . . . . . . . 503
Mojtaba Heidarysafa, James Reed, Kamran Kowsari,
April Celeste R. Leviton, Janet I. Warren, and Donald E. Brown
3D Reconstruction Under Weak Illumination Using
Visibility-Enhanced LDR Imagery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515
Nader H. Aldeeb and Olaf Hellwich
DynFace: A Multi-label, Dynamic-Margin-Softmax Face
Recognition Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535
Marius Cordea, Bogdan Ionescu, Cristian Gadea, and Dan Ionescu
Towards Resolving the Kidnapped Robot Problem: Topological
Localization from Crowdsourcing and Georeferenced Images . . . . . . . . 551
Sotirios Diamantas
Using the Z-bellSM Test to Remediate Spatial Deficiencies
in Non-Image-Forming Retinal Processing . . . . . . . . . . . . . . . . . . . . . . . 564
Clark Elliott, Cynthia Putnam, Deborah Zelinsky, Daniel Spinner,
Silpa Vipparti, and Abhinit Parelkar
Learning of Shape Models from Exemplars of Biological Objects
in Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 580
Petra Perner
A New Technique for Laser Spot Detection and Tracking
by Using Optical Flow and Kalman Filter . . . . . . . . . . . . . . . . . . . . . . . 600
Xiuli Wang, Ming Yang, Lalit Gupta, and Yang Bai
Historical Document Image Binarization Based on Edge
Contrast Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614
Zhenjiang Li, Weilan Wang, and Zhengqi Cai
Development and Laboratory Testing of a Multipoint Displacement
Monitoring System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 629
Darragh Lydon, Su Taylor, Des Robinson, Necati Catbas, and Myra Lydon
Contents xi
Quantitative Comparison of White Matter Segmentation

for Brain MR Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
Xianping Li and Jorgue Martinez
Evaluating the Implementation of Deep Learning in LibreHealth
Radiology on Chest X-Rays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 648
Saptarshi Purkayastha, Surendra Babu Buddi, Siddhartha Nuthakki,
Bhawana Yadav, and Judy W. Gichoya
Illumination-Invariant Face Recognition by Fusing Thermal
and Visual Images via Gradient Transfer . . . . . . . . . . . . . . . . . . . . . . . 658
Sumit Agarwal, Harshit S. Sikchi, Suparna Rooj,
Shubhobrata Bhattacharya, and Aurobinda Routray
An Attention-Based CNN for ECG Classification . . . . . . . . . . . . . . . . . . 671
Alexander Kuvaev and Roman Khudorozhkov
Reverse Engineering of Generic Shapes Using Quadratic Spline
and Genetic Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 678
Misbah Irshad, Munazza Azam, Muhammad Sarfraz,
and Malik Zawwar Hussain
Bayesian Estimation for Fast Sequential Diffeomorphic
Image Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687
Youshan Zhang
Copyright Protection and Content Authentication Based on
Linear Cellular Automata Watermarking for 2D Vector Maps . . . . . . . 700
Saleh AL-ardhi, Vijey Thayananthan, and Abdullah Basuhail
Adapting Treemaps to Student Academic Performance Visualization . . . 720
Samira Keivanpour
Systematic Mobile Device Usage Behavior and Successful
Implementation of TPACK Based on University Students Need . . . . . . . 729
Syed Far Abid Hossain, Yang Ying, and Swapan Kumar Saha
Data Analysis of Tourists’ Online Reviews on Restaurants
in a Chinese Website . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 747
Meng Jiajia and Gee-Woo Bock
Body of Knowledge Model and Linked Data Applied
in Development of Higher Education Curriculum . . . . . . . . . . . . . . . . . 758
Pablo Alejandro Quezada-Sarmiento, Liliana Enciso, Lorena Conde,
Monica Patricia Mayorga-Diaz, Martha Elizabeth Guaigua-Vizcaino,
Wilmar Hernandez, and Hironori Washizaki
Building Adaptive Industry Cartridges Using a Semi-supervised
Machine Learning Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 774
Lucia Larise Stavarache
xii Contents
Decision Making with Linguistic Information for the Development

of New Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 789
Zapata C. Santiago, Escobar R. Luis, and Ponce N. Alvaro
Researcher Profile Ontology for Academic Environment . . . . . . . . . . . . 799
Maricela Bravo, José A. Reyes-Ortiz, and Isabel Cruz
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 819
Deep Learning for Detection of Railway
Signs and Signals
Georgios Karagiannis1,2(B) , Søren Olsen1 , and Kim Pedersen1

1
Department of Computer Science, University of Copenhagen,
2200 Copenhagen, Denmark
{geka,ingvor,kimstp}@di.ku.dk
2
COWI A/S, Parallelvej 2, 2800 Lyngby, Denmark
geks@cowi.com
Abstract. Major railway lines need advance management systems

based on accurate maps of their infrastructure. Asset detection is an
important tool towards automation of processes and improved decision
support on such systems. Due to lack of available data, limited research
exists investigating railway asset detection, despite the rise of Artifi-
cial Neural Networks and the numerous investigations on autonomous
driving. Here, we present a novel dataset used in real world projects for
mapping railway assets. Also, we implement Faster R-CNN, a state of
the art deep learning object detection method, for detection of signs and
signals on this dataset. We achieved 79.36% on detection and a 70.9%
mAP. The results were compromised by the small size of the objects, the
low resolution of the images and the high similarity across classes.
Keywords: Railway · Object detection · Object recognition ·

Deep learning · Faster R-CNN
1 Introduction
The ever increasing modernisation of signal systems and electrification of major
railway lines lead to increasingly complex railway environments. These environ-
ments require advanced management systems which incorporate a detailed rep-
resentation of the network, its assets and surroundings. The aim of such systems
is to facilitate automation of processes and improved decision support, minimis-
ing the requirement for expensive and inefficient on-site activities. Fundamental
requirements are detailed maps and databases of railway assets, such as poles,
signs, wires, switches, cabinets, signalling equipment, as well as the surrounding
environment including trees, buildings and adjacent infrastructure. The detailed
maps may also form the basis for a simulation of the railway as seen from the
train operator’s viewpoint. Such simulations/videos are used in the training of
train operators and support personnel. Ideally, the maps should be constantly
updated to ensure currency of the databases as well as to facilitate detailed
documentation and support of maintenance and construction processes in the
c Springer Nature Switzerland AG 2020
K. Arai and S. Kapoor (Eds.): CVC 2019, AISC 943, pp. 1–15, 2020.
https://doi.org/10.1007/978-3-030-17795-9_1
2 G. Karagiannis et al.
networks. However, with currently available methods, mapping of railway assets

is largely a manual and highly labour intensive process, limiting the possible
levels of detail and revisit times. The response to this challenge is to automate
railway asset mapping based on different sensor modalities (2D images or 3D
point clouds) acquired from ground or air.
Despite the high demand for automatic asset detection along railways, there
is very little research on this field [1]. Here, we present an approach on detection
of signs and signals as a first step towards automatic generation and update of
maps of railway environments. We implement an object detection model based
on Faster R-CNN (Region-based Convolutional Neural Network) presented by
Ren et al. in [2] on a dataset used to map a railway of 1,700 km in 2015. The
mapping was carried out manually from a private company1 with a lot of expe-
rience in such projects. Currently, many such projects exist around the world
and to the best of our knowledge are still carried out manually (people going
through all images and mark objects of interest). Our approach aims to show the
performance of an advanced object detection algorithm, such as Faster R-CNN,
on a novel dataset used in a real world project.
2 Literature Review
2.1 Previous Approaches
The research on automatic object detection along railways is sparse, compared

to the analogous, popular field of road furniture detection, mainly due to the
lack of available railway traffic data [1]. Most of the research is focused on pas-
senger detection [3,4] or track detection [5–9] for different purposes. The limited
research existing focused on detection of only a single type of object (sign recog-
nition [10], sign detection [11] or wire detection [12]).
Marmo et al. [10] presented a classical approach for railway signal detection.
It is focused on detecting a specific type of signals (single element) in video
frames and classify it according to the colour of the light (green-pass, red-no
pass). The implementation is based on simple image processing techniques such
as histogram analysis, template matching and shape feature extraction. The
method resulted in 96% detection accuracy and 97% classification accuracy in a
total of 955 images which is impressing for this type of approaches. The advan-
tage of this method is efficiency, however it is focused on a very specific type of
signals and the examples presented are scenes of low complexity.
Arastounia [12] presented an approach for detection of railway infrastructure
using 3D LiDAR data. The approach is focused on detection of cables related to
the railway (i.e. catenary, contact or return current cables), track bed, rail tracks
and masts. The data covers about 550 m of Austrian railways. The approach is
based mainly on the topology of the objects and their spatial properties. Points
on the track bed are first detected from a spatially local statistical analysis. All
the other objects of interest are recognised depending on their spatial relation
1
Second Affiliation.
Railway Signs and Signals 3
with the track bed. The overall average detection achieved by this method is
96.4%. The main drawback of this approach is that it depends on a sophisticated
type of data that needs special equipment to capture and it is more complicated
to process.
Agudo et al. in [1] presented a real-time railway speed limit and warning
signs recognition method on videos. After noise removal, Canny edge detection
is applied and an optimised Hough voting scheme detects the sign region of
interest on the edge image. Based on gradient directions and distances of points
on the edge of a shape, candidate central points of signs occur. Then, recognition
is achieved by applying shape criteria since the signs are either circular, square
or rectangular. The method scored 95.83% overall accuracy on classification.
However, even though the dataset had more than 300,000 video frames, only
382 ground truth signs existed and the authors do not provide any score for the
detection accuracy.
2.2 Convolutional Neural Networks

All of the above approaches are using traditional image analysis methods to
solve object detection problems. To the best of our knowledge, there are no
published methods that attempt to solve object detection in railways based on
Convolutional Neural Networks (CNNs). CNNs represent the state of the art
concept in Computer Vision for classification, detection and semantic segmenta-
tion. Regarding object detection, we can divide most CNN-based methods into
two categories: region-based and single shot methods. The most characteristic
representative of the first is Region-based CNN (R-CNN) and its descendants
Fast, Faster and the recent Mask R-CNNs [2,13–15]. From the second category,
most representative methods are You Only Look Once (YOLO) [16] and Sin-
gle Shot multibox Detector (SSD) [17]. In general, region based methods are
considerably slower but more effective. Also, region based methods show better
performance on smaller objects [2,16]. Given the performance shown in com-
petitive challenges and the fact that our dataset consists mainly of very small
objects, we Faster R-CNN [2] consider more suitable for our problem.
3 Data Analysis
3.1 Dataset
In our case, the dataset consists of 47,912 images acquired in 2013 and show the
railway from Brisbane to Melbourne in Australia. The images were acquired in
three different days in the morning with similar sunlight conditions. The camera
used is a Ladybug3 spherical camera system and the images are panoramic views
of size 5400 × 2700 pixels. The images were processed manually by the produc-
tion team of the company2 for annotations and resulted in 121,528 instances of
railway signs and signals.
2
Second Affiliation.
Fig. 1. Instances of signals and signs. First row from left to right (class name in paren-
theses when it is different from the description): speed sign (Sign S), unit sign (Sign
U), speed standard letter (Sign Other), position light main (PL 2W), position light
separately (PL 2W& 1R), signal with two elements (Signal2 F). Second row: diverging
left-right speed sign (Sign LR), diverging left speed sign (Sign L), signal number (Sign
Signal), signal with one element (Signal1 F), signal with three elements (Signal3 F),
signal with four elements (Signal4 F).
Fig. 2. Instances of speed signs at different lighting conditions, viewpoints and scales.
The samples are originally separated into twenty five classes. Each class is in
fact a subclass of two parent classes, signs and signals. Fifteen classes correspond
to signals: three different types of position lights with their back and side view,
signals with one, two, three or four elements (lights), their side and back view
and other type of signals. Also, ten classes correspond to different types of signs:
speed signs, diverging left speed signs, diverging right speed signs, diverging
left-right speed signs, unit signs, signal number signs, other signs, back views of
circular signs, back views of diamond signs and back views of rectangular signs.
From the total amount of samples, 67,839 correspond to signals and 53,689 to
signs. Figure 1 shows some instances of signs and signals. Each of these instances
correspond to a different class. We can see the high similarity among the classes
both for signs and signals. Specifically, diverging left, diverging right, diverging
left-right and regular speed signs are very similar and especially when they are
smaller than the examples shown in Fig. 1. Similarly, even for humans it is often
hard to distinguish between signals with three or four elements when they are
small. All examples shown here are larger than their average size on the dataset
for clarity.
Figure 2 shows examples of regular speed signs with different viewpoint, size
and illumination. These examples illustrate the scale, viewpoint and illumination
variation of the objects but at the same time the high similarity among the
classes.
Fig. 3. Instances of signals with one element (Signal1). All four examples belong to
the same class even though they have different characteristics. From left to right: some
have long (first and second), no top cover at all (third) or short cover (last). Also, some
have no back cover (first), other have circular (second and third) or semicircular (last)
Figure 3 shows examples of the same class of signals (Signal1). It is important

to note that the class is one of the least represented in the dataset with only a
few hundred samples available. Though, despite the low availability we can see
that there is significant intra-class variability. The same is observed in all the
other classes of signals except the ones corresponding to position lights.
Figure 4 shows the amount of available samples for each class. These quan-
tities vary widely for the different classes. For instance, there are about 23,000
samples available for the front view of 4-lamp signals but only a few hundreds
for position lights with two lamps or for diverging left-right speed signs. Our
dataset reflects the real distribution of objects along a railway, which means
that in a railway there exist very few position lights with two lamps and diverg-
ing left-right speed signs. Therefore, this level of imbalance among the classes
is unavoidable in real applications. However, in deep learning, large amounts of
samples are necessary to train a robust model.
A common workaround to ensure adequate balance among classes is to apply
some data augmentation techniques (e.g. apply random crops, rotations, scal-
ing, illumination changes etc. on the existing samples). However, such tech-
niques cannot solve the problem, without causing bias to the dataset, in our
case because the difference in available samples is too high. Past observations
[18], have found that non-iconic samples may be included during training only
if the overall amount of samples is large enough to capture such variability. In
any other case, these samples may act as noise and pollute the model. Thus,
it is necessary for some classes with similar characteristics to be merged. By
merging all signs except speed and signal number signs to a general class signs
other, we get a class with about 30,000 samples. Also, we merged all position
lights to a single class, resulting in about 10,000 samples for this class. Finally,
the front and side views of each signal class were merged into a single class. The
back views of all signals remained a separate class because there was no specific
information available on which type of signal each back view belonged to. After
these operations, we end up with ten classes of at least a few thousand samples
each (Fig. 5). This way, the misrepresentation problem is softened, however we
introduce high intra-class variability. The least represented class is the single
element signals (Signal 1) with about 6,000 samples which is still about four
times less than the most dominant class, but more manageable.
Fig. 4. Amount of sample instances per class before merging some classes. The imbal-
ance among the classes is too high.
Fig. 5. Amount of sample instances per class. After merging some classes with similar
characteristics we end up with a more balanced dataset.
Another important aspect of the samples is their size. Figure 6 is a histogram

of the size of the samples in square pixels. About 65% of the samples have area
less than 1000 pixels (≈322 ) and 89% less than 2500 pixels (502 ). Given the
size of the panoramic images, a 502 sample corresponds to 0.018% of the whole
image. In COCO [18], one of the most challenging datasets in Computer Vision,
the smallest objects correspond to 4% of the entire image. The winners for the
2017 COCO Detection: Bounding Box challenge achieved less than 55% accuracy.
This is an indication of the difficulty of our problem, in terms of relative size of
objects.
Fig. 6. Amount of samples according to their size in pixel2 . Most samples (89%) are
smaller than 502 pixels.
A reason behind the high amount of very small objects in our dataset is that
the data was acquired by driving on a single track. However, in many sectors
along the railway there are multiple parallel tracks and the signs and signals
corresponding to these tracks appear in the dataset only in small sizes since
the camera never passed close to them. One way to limit the small object size
problem in our dataset is to split each panoramic image into smaller patches with
high overlap, small enough to achieve a less challenging relative size between
objects and image. Specifically, in our approach, each image is split into 74
patches of size 6002 pixels with 200 pixels overlap on each side. Even at this
level of fragmentation, a 502 object corresponds to 0.69% of a patch size. A
consequence of splitting the images into smaller patches is that the same object
may now appear in more than one patches due to overlap. In fact, while on the
panoramic images there exist 121,528 object instances, on the patches that were
extracted, there exist 203,322 instances. The numbers shown in Fig. 6 correspond
to the objects existing on the patches.
3.2 Railway vs Road Signs and Signals
Here, it is important to highlight the difference between the problem described in

this paper and the more popular topic of the detection of road signs and signals.
The most important difference is the size of the objects. The height of a road
sign varies from 60 to 150 cm depending on the type of road [19] while in railways
it is usually less than 40 cm high [20]. Given also that most of the times the signs
are located only a few centimetres from the ground supported by a very short
pole, it much harder to detect. Also, in railways, signs are often very similar but
have different meaning like the first two examples of the second row in Fig. 1.
At the same time, it is very common along the same railway objects of the same
class to look different as shown in Fig. 3. Thus, a detector of railway signs and
signals needs to be able to distinguish objects based on fine details. Finally, in

railways the signs and the signals are often combined creating more complex
structures posing an extra challenge on a detection algorithm (e.g. the detected
signals shown in Fig. 10). Given the above differences, we consider railway signs
a more challenging detection problem.
4 Methodology
4.1 Faster R-CNN
For the detection of signs and signals, we applied the Faster R-CNN presented
by Ren et al. in [2] using ResNet-101 [21] as feature extractor. We decided to
implement this approach mainly motivated by its high performance on competi-
tive datasets such as Pascal VOC 2007 (85.6% mAP), Pascal VOC 2012 (83.8%
mAP) and COCO (59% mAP). The main drawback of this approach compared
to other acknowledged object detection methods such as YOLO [16]) or SSD is
its high processing time (three to nine times slower depending on the implemen-
tation [17]). However, the sacrifice in time pays off in accuracy, especially in this
dataset since this method performs better on small objects [2].
Here we will present some key points of Faster R-CNN. First, Faster R-
CNN is the descendant of Fast R-CNN [2] which in turn is the descendant of
R-CNN [13]. As their names imply, Fast and Faster R-CNNs are more efficient
implementations of the original concept in [13], R-CNN. The main elements of
Faster R-CNN are: (1) the base network, (2) the anchors, (3) the Region Proposal
Network (RPN) and (4) the Region based Convolutional Neural Network (R-
CNN). The last element is actually Fast R-CNN, so with a slight simplification
we can state that Faster R-CNN = RPN + Fast R-CNN.
The base network is a, usually deep, CNN. This network consists of multi-
ple convolutional layers that perform feature extraction by applying filters at
different levels. A common practice [2,14,16] is to initialize training using a
pre-trained network as a base network. This helps the network to have a more
realistic starting point compared to random initialization. Here we use ResNet
[21]. The second key point of this method is the anchors, a set of predefined
possible bounding boxes at different sizes and aspect ratios. The goal of using
the anchors is to catch the variability of scales and sizes of objects in the images.
Here we used nine anchors consisting of three different sizes (152 , 302 and 602
pixels) and three different aspect ratios 1:1, 1:2 and 2:1.
Next, we use RPN, a network trained to separate the anchors into fore-
ground and background given the Intersection over Union (IoU) ratio between
the anchors and a ground-truth bounding box (foreground if IoU > 0.7 and
background if IoU < 0.1). Thus, only the most relevant anchors for our dataset
are used. It accepts as input the feature map output of the base model and
creates two outputs: a 2 × 9 box-classification layer containing the foreground
and background probability for each of the nine different anchors and a 4 × 9
box-regression layer containing the offset values on x and y axis of the anchor
Fig. 7. Right: the architecture of Faster R-CNN. Left: the region proposal network
(RPN). Source: Figs. 2 and 3 of [2].
bounding box compared to the ground-truth bounding boxes. To reduce redun-

dancy, due to overlapping bounding boxes, non-maximum suppression is used
on the proposed bounding boxes based on their score on the box-classification
output. A threshold of 0.7 on the IoU is used resulting in about 1,800 proposal
regions per image in our case (about 2,000 in the original paper).
Afterwards, for every region proposal we apply max pooling on the features
extracted from the last layer of the base network. Finally, the Fast RCNN is
implemented, mainly by two fully-connected layers as described originally in
[14]. This network outputs a 1 × (N + 1) vector (a probability for each of the N
number of classes plus one for the background class) and a 4 × N matrix (where
4 corresponds to the bounding box offsets across x and y axis and N on the
number of classes). Figure 7 shows the structure of Faster R-CNN and the RPN.
The RPN and R-CNN are trained according to the 4-step alternate training
to learn shared features. At first, the RPN is initialized with ResNet and fine
tuned on our data. Then, the region proposals are used to train the R-CNN
separately, again initialized by the pre-trained ResNet. Afterwards, the RPN is
initialized by the trained R-CNN with the shared convolutional layers fixed and
the non-shared layers of RPN are fine tuned. Finally, with the shared layers
fixed, the non-shared layers of R-CNN are fine tuned. Thus, the two networks
are unified.
4.2 Evaluation Method
For the evaluation of detection, an overlap criterion between the ground truth
and predicted bounding box is defined. If the Intersection over Union (IoU) of
these two boxes is greater than 0.5, the prediction is considered as True Positive
(TP). Multiple detections of the same ground truth objects are not considered
true positives, each predicted box is either True-Positive or False-Positive (FP).
Predictions with IoU smaller than 0.5 are ignored and count as False Negatives
(FN). Precision is defined as the fraction of the correct detections over the total
detections ( T PT+F
P
P ). Recall is the fraction of correct detections over the total
amount of ground truth objects ( T PT+F P
N ) [22]. For the evaluation of classifica-
tion and overall accuracy of our approach, we adopted mean Average Precision
(mAP) [23] as a widely accepted metric [22]. For each class, the predictions
satisfying the overlap criterion are assigned to ground truth objects in descend-
ing order by the confidence output. The precision/recall curve is computed and
the average precision(AP) is the mean value of interpolated precision at eleven
equally spaced levels of recall [22]:
1
AP = pinterp (r) (1)
11
r∈{0,0.1,...,1}
where
pinterp (r) = maxr̃:r̃≥r p(r̃) (2)
Then, the mean of all APs across the classes is the mAP metric. This metric
was used as evaluation method for the Pascal VOC 2007 detection challenge
and from that time is the most common evaluation metric on object detection
challenges [18].
5 Results
The network was trained on a single Titan-X GPU implementation for approx-
imately two days. We trained it for 200k iterations with a learning rate of 0.003
Fig. 8. Percentage of objects detected according to their size in pixels2 . The algorithm
performed very well on detecting objects larger than 400 pixels (more than 79% in
all size groups). On the other hand, it performed poorly on very small objects (less
than 200 pixels) detecting only 24% of them. The performance was two times better
on objects with size 200–400 pixels (57%), which is the most dominant size group with
about 45,000 samples (see Fig. 6).
and for 60k iterations with learning rate 0.0003. The model achieved an overall
precision of 79.36% on the detection task and a 70.9% mAP.
The precision level of detection is considered high given the challenges
imposed by the dataset as they are presented in Sect. 3. Figure 8 shows the
detection performance of the algorithm with respect to the size of the samples.
We can see that the algorithm fails to detect very small objects. About 76%
of objects smaller than 200 pixels are not detected. A significantly better, but
still low, performance is observed for objects of 200–400 pixels size (57% detec-
tion rate). On the other hand, the algorithm performs uniformly well for objects
Fig. 9. Example of detection. Two successful detections and one false positive. We
can see the illumination variance even in a single image with areas in sunlight and in
shadow. The sign is detected successfully despite its small size and the partial occlusion
from the bush. The entrance of the tunnel was falsely detected as a signal with three
elements.
larger than 400 pixels (79% for sizes 400–600 and more than 83% for larger than
600 pixels). These results show that, in terms of object size, there is a threshold
above which the performance of Faster-RCNN is stable and it is unaffected by
the size of the objects. Therefore, if there are enough instances and the objects
are large enough, object size is not a crucial factor. In our case, this threshold
is about 500 pixels.
An interesting metric for our problem would be to measure the amount of
unique physical objects detected. The goal of object detection in computer vision,
usually, is to detect all the objects that appear in an image. In mapping, most
of the times, it is important to detect the locations of the physical objects. If we
have multiple images showing the same physical object from different point of
views and at different scales, in mapping, it would be sufficient to detect it at
Fig. 10. Example of detection. A successful detection of more complex object struc-
tures. Different signals mounted on top of each other did not confuse the algorithm.
least in one image and not necessarily in all images. We expect that our approach
will perform better in this metric, however, information about unique physical
objects is not available for this dataset.
Figures 9 and 10 show two representative examples of the results. We can see
in these figures the variance in illumination conditions and the small size of the
objects. The speed signs that appear on these images are very small but they
belong to the most common object size interval for this dataset (200–400 pixels).
By looking at these figures, we can also realise the difficulty of the classification
task of this dataset. Even for a human, it is hard to decide on which class the
objects on these images belong to.
Fig. 11. Confusion matrix showing the classification AP of the detected objects. Rows
represent the predictions and columns the ground truth objects. Last row and last
column (red colour) summarise the performance per class. Bottom right cell (blue
colour) is the mAP of the method. See Fig. 2 for class name explanation.
Figure 11 is a confusion matrix that summarises the performance of the

method on classification. The diagonal represents the Average Precision (as
described in Sect. 4.2) for each class and the bottom right cell is the mean
of these precisions. A first observation is that there is significant variation in
accuracy for the different classes. As expected, the accuracy is decreased for the
classes trained with few samples (position lights) and for those with high simi-
larities with other classes (back view of signals). An interesting exception to this
observation is the class signals with one element (S1). Despite the low amount
of training samples and the fact that all classes of signals show high resemblance
among each other, the classification accuracy is above the overall and among the
highest.
6 Conclusions
In this paper, we presented a novel dataset used in real world projects for map-
ping of railway infrastructure. The dataset is very challenging mainly because
the objects are very small relatively to the size of the entire images but also
absolutely in terms of pixels2 . We implemented a state of the art deep learn-
ing method for detection and classification of signs and signals on this dataset
scoring 79.36% on the detection task and 70.9% on the classification task. We
believe that the accuracy of the model will be higher in a similar dataset with
better resolution images. Also, given that the object positions along a railway fol-
low specific regulations, taking into account the spatial relationships among the
objects can improve accuracy. Moreover, it would be interesting to apply other
state of the art methods on this dataset and analyse the strengths and weak-
nesses of each method on the same tasks. In addition, for mapping tasks such
as the current, a more appropriate metric would be to evaluate detections based
on the amount of physical objects and not on objects existing in the images. In
conclusion, our results are promising and suggest further investigation into the
use of deep learning for railway asset detection and mapping.
References
1. Agudo, D., Sánchez, Á., Vélez, J.F., Moreno, A.B.: Real-time railway speed limit
sign recognition from video sequences. In: 2016 International Conference on Sys-
tems, Signals and Image Processing (IWSSIP), pp. 1–4, May 2016
2. Ren, S., He, K., Girshick, R., Sun, J.: Faster R-CNN: towards real-time object
detection with region proposal networks. In: Advances in Neural Information Pro-
cessing Systems, vol. 2015, pp. 91–99 (2015)
3. Zheng, D., Wang, Y.: Application of an artificial neural network on railway passen-
ger flow prediction. In: Proceedings of 2011 International Conference on Electronic
Mechanical Engineering and Information Technology, vol. 1, pp. 149–152, August
2011
4. Tsai, T.-H., Lee, C.-K., Wei, C.-H.: Neural network based temporal feature mod-
els for short-term railway passenger demand forecasting. Expert Syst. Appl.
36(2), 3728–3736 (2009). Part 2. http://www.sciencedirect.com/science/article/
pii/S0957417408001516
5. Karthick, N., Nagarajan, R., Suresh, S., Prabhu, R.: Implementation of railway
track crack detection and protection. Int. J. Eng. Comput. Sci. 6(5), 21476–21481
(2017). http://ijecs.in/index.php/ijecs/article/view/3535
6. Sadeghi, J., Askarinejad, H.: Application of neural networks in evaluation of railway
track quality condition. J. Mech. Sci. Technol. 26(1), 113–122 (2012). https://doi.
org/10.1007/s12206-011-1016-5
7. Gibert, X., Patel, V.M., Chellappa, R.: Robust fastener detection for autonomous
visual railway track inspection. In: 2015 IEEE Winter Conference on Applications
of Computer Vision, pp. 694–701, January 2015
8. Sinha, D., Feroz, F.: Obstacle detection on railway tracks using vibration sensors
and signal filtering using bayesian analysis. IEEE Sens. J. 16(3), 642–649 (2016)
9. Faghih-Roohi, S., Hajizadeh, S., Núñez, A., Babuska, R., Schutter, B.D.: Deep
convolutional neural networks for detection of rail surface defects. In: 2016 Inter-
national Joint Conference on Neural Networks (IJCNN), pp. 2584–2589, July 2016
10. Marmo, R., Lombardi, L., Gagliardi, N.: Railway sign detection and classifica-
tion. In: 2006 IEEE Intelligent Transportation Systems Conference, pp. 1358–1363,
September 2006
11. Melander, M., Halme, I.: Computer vision based solution for sign detection.
Eur. Railway Rev. (2016). https://www.globalrailwayreview.com/article/30202/
computer-vision-based-solution-sign-detection/
12. Arastounia, M.: Automated recognition of railroad infrastructure in rural areas
from LIDAR data. Remote Sens. 11, 14916–14938 (2015)
13. Girshick, R., Donahue, J., Darrell, T., Malik, J.: Rich feature hierarchies for accu-
rate object detection and semantic segmentation. In: IEEE Conference on Com-
puter Vision and Pattern Recognition (CVPR), pp. 580–587 (2014)
14. Girshick, R.: Fast R-CNN. In: Proceedings (IEEE International Conference on
Computer Vision), vol. 2015, pp. 1440–1448 (2015)
15. He, K., Gkioxari, G., Dollár, P., Girshick, R.: Mask R-CNN. In: Proceedings of the
International Conference on Computer Vision (ICCV) (2017)
16. Redmon, J., Divvala, S., Girshick, R., Farhadi, A.: You only look once: unified,
real-time object detection. In: IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), vol. 2016, pp. 779–788 (2016)
17. Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., Berg, A.C.:
SSD: Single shot multibox detector. In: Lecture Notes in Computer Science, vol.
9905, pp. 21–37 (2016)
18. Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollar, P.,
Zitnick, C.L.: Microsoft COCO: common objects in context. In: Lecture Notes in
Computer Science, vol. 8693, pp. 740–755 (2014)
19. Great Britain: Department for Transport. Traffic Signs Manual. The Stationary
Office, London, United Kingdom (2013)
20. Safety, R., Board, S.: Lineside Operational Safety Signs. Railway Group Standard,
London (2009)
21. He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition.
In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), vol.
2016, pp. 770–778 (2016)
22. Everingham, M., Gool, L., Williams, C.K., Winn, J., Zisserman, A.: The pascal
visual object classes (VOC) challenge. Int. J. Comput. Vis. 88(2), 303–338 (2010).
https://doi.org/10.1007/s11263-009-0275-4
23. Salton, G., McGill, M.J.: Introduction to Modern Information Retrieval. McGraw-
Hill Inc., New York (1986)
3D Conceptual Design Using Deep Learning
Zhangsihao Yang, Haoliang Jiang, and Lan Zou(&)
Carnegie Mellon University, Pittsburgh, PA 15213, USA

{zhangsiy,haolianj,lzou1}@andrew.cmu.edu
Abstract. This article proposes a data-driven methodology to achieve fast

design support to generate or develop novel designs covering multiple object
categories. This methodology implements two state-of-the-art Variational
Autoencoder, dealing with 3D model data, with a self-defined loss function. The
loss function, containing the outputs of individual layers in the autoencoder,
obtains combinations of different latent features from different 3D model cate-
gories. This article provides detail explanation for utilizing the Princeton Model-
Net40 database, a comprehensive clean collection of 3D CAD models for
objects. After converting the original 3D mesh file to voxel and point cloud data
type, the model will feed an autoencoder with data in the same dimension. The
novelty is to leverage the power of deep learning methods as an efficient latent
feature extractor to explore unknown designing areas. The output is expected to
show a clear and smooth interpretation of the model from different categories to
generate new shapes. This article will explore (1) the theoretical ideas, (2) the
progress to implement Variational Autoencoder to attain implicit features from
input shapes, (3) the results of output shapes during training in selected domains
of both 3D voxel data and 3D point cloud data, and (4) the conclusion and future
work to achieve the more outstanding goal.
Keywords: Design support Data analysis 3D representation

Generative model Computer vision
1 Introduction
Three-dimensional (3D) representations of real-life objects have been an essential

application in computer vision, robotics, medicine, virtual reality, and mechanical
engineering. Due to the increase of interest in tapping into latent information of 3D
models in both academic and industrial areas, a generative neural network of 3D
objects has achieved impressive improvement in the last several years. Recent research
achievement on generative deep learning neural network concepts includes autoen-
coder [4], Variational Autoencoder (VAE) [1] and generative adversarial model [6].
These models have led great progress in natural language processing [3], translate
machine and image generation [10]. With the capability of extract features from data,
these models are also widely used in generating embedding vectors [10]. In studies of
3D shapes, researchers utilized these state-of-the-art generative neural networks to take
latent feature vectors as the representation of 3D voxelized shapes and 3D point cloud
shapes. Those researches show promising results during experiments. To support the
© Springer Nature Switzerland AG 2020

K. Arai and S. Kapoor (Eds.): CVC 2019, AISC 943, pp. 16–26, 2020.
https://doi.org/10.1007/978-3-030-17795-9_2
Another random document with
no related content on Scribd:
However, from whatever cause arising, we had the full benefit of the
peal.
“We got to our rural retreat about half-past nine, both the men-
servants attending us on horseback. At ten o’clock your old Mumpsa
[the child’s pet name for her mother] laid the first brick, and placed
under it a medal struck in commemoration of the centenary of the
Revolution of 1688. Your darling father then placed another for
himself, and a third for his beloved treasure, which he made Toney
put her foot upon; and after the little rogue had done so, you would
have laughed to have seen how she wagged her tail, and nodded
her head upon it, as much as to say she was very proud of being
admitted to have, not a finger, but a foot, in the business. The men
worked merrily on until two o’clock, and then repaired to the public-
house, where two legs of mutton, and bread, beer, and potatoes
were provided for them. There they enjoyed themselves for the rest
of the day, and this morning cheerily resumed their labours.”
Having thus impressed the natives, including the landlord of “The
Bell,” with a sense of the importance of the new owner about to
come among them, Dr. Mitford completed the business by
substituting the name “Bertram House” for that of “Grazeley Court,”
the reason for which, did the curiosity of the neighbouring aristocracy
cause them to inquire, was to be discovered in the fact that he was a
scion of the Mitfords of Bertram Castle, Bertram being the original
and ancient name of the family.
Judging by the very scant records of this period at our command, it
would appear that the erection of Bertram House, and its completion
to Dr. Mitford’s satisfaction, must have occupied nearly four years.
This would give Miss Mitford a clear three years of life among the
mild excitements which Reading then offered before taking up
residence at Grazeley in a district which she was to immortalize—the
term surely needs no justification for its use—and in which she was
destined, save for a few notable occasions when duty or
considerations of health called her away for short periods, to live out
her life to the end.
Her introduction to the gaieties of this respectable Borough took
place in the August of 1803, when she would be nearly sixteen. The
occasion was the annual Race Ball, at which function it was the time-
honoured custom of the race-steward to dance with the young ladies
then making their début, an ordeal almost as trying to the débutante
in those prim and decorous days as a presentation at Court,
especially if the steward happened to be a total stranger to her.
Writing to her mother, towards the end of her school career and
commenting on this, Miss Mitford added—possibly to gain courage
from the inditing—“I think myself very fortunate that Mr. Shaw
Lefevre[8] will be steward next year, for by that time I shall hope to
know him well enough to render the undertaking of dancing with him
much less disagreeable.” In this connexion we venture to suggest
that on this occasion Mr. Shaw Lefevre would have full hands, when
we remember that even at this comparatively early date Miss
Mitford’s figure had already assumed generous proportions and that
she was short of stature into the bargain.
Naturally enough, the conclusion of school life and the re-
commencement of life at home afforded the young girl the fullest
opportunities for observing, noting and commenting on persons and
events, a pastime in which she delighted. Her pictures of the
Reading of her day are notable alike for their quaint fancies as for
their fidelity. Her picture of the town—which she disguises under the
name of Belford Regis—as viewed from the southern heights of
Whitley, is one to which all true lovers of the old town turn with
pleasure even to-day.
“About this point,” she says, “is perhaps to be seen the very best
view of Belford, with its long ranges of modern buildings in the
outskirts, mingled with picturesque old streets; the venerable towers
of St. Stephen’s [St. Mary’s] and St. Nicholas [St. Laurence’s]; the
light and tapering spire of St. John’s [St. Giles’]; the huge monastic
ruins of the abbey; the massive walls of the county gaol; the great
river winding along like a thread of silver; trees and gardens mingling
amongst all; and the whole landscape enriched and lightened by the
dropping elms of the foreground, adding an elusive beauty to the
picture, by breaking the too formal outline and veiling just exactly
those parts which most require concealment. Nobody can look at
Belford from this point, without feeling that it is a very English and
very charming scene; and the impression does not diminish on
further acquaintance.”
Continuing, she compares the old romantic structures in which our
ancestors delighted—now, unhappily, nearly all demolished—with,
what she calls, the handsome and uniform buildings which are now
the fashion; and she remarks on the rapid growth which the town
was then making, “having recently been extended to nearly double
its former size.” What would she have said, we wonder, could she
have foreseen the Reading of to-day with its palatial polished-
granite-fronted business emporiums controlled from the Metropolis
by great limited liability companies whose insatiable appetites are
devouring, as their policy of grab is choking, the life from the old-time
burgesses; burgesses who gloried in their town and whom their town
took pleasure in honouring; men whose places are now filled by
battalions of shopmen whose fixity of tenure is so doubtful as to
preclude them from taking any part or interest, however slight, in the
town which shelters them? And, in regard to the extension which she
names with so much pride, how she would gasp with astonishment
had she been told that Whitley, from which she viewed the pleasant
scene, would be turned into dreary streets of uniformly built villas,
never deviating by so much as half a brick from the monotony of the
usual “desirable residence”; that the old limits of the town, beyond
which she could easily descry the panoramic revel of field and
meadow, would be extended for nearly two miles each way, almost
indeed to her beloved “Our Village,” and that the population of
16,000—each unit placidly pursuing its fairly prosperous calling—
would be transformed, seventy years later, into a struggling,
perspiring, more or less harassed army of 88,000, the majority not
daring, though they would not admit the stern impeachment, to call
their bodies their own.
“The good town of Belford,” she later remarks, “swarmed of course
with single ladies ... and was the paradise of ill-jointured widows and
portionless old maids. They met on the tableland of gentility, passing
their mornings in calls at each other’s houses, and their evenings in
small tea-parties, seasoned with a rubber or a pool, and garnished
with the little quiet gossiping (call it not scandal, gentle reader!)
which their habits required.... The part of the town in which they
chiefly congregated, the lady’s quartier, was one hilly corner of the
parish of St. Nicholas, a sort of highland district, all made up of short
rows, and pigmy places, and half-finished crescents, entirely
uncontaminated by the vulgarity of shops,” chosen, it is suggested,
“perhaps because it was cheap, perhaps because it was genteel—
perhaps from a mixture of both causes.” A kindly satire this, and
interesting because it points so conclusively to a certain backwater
near the Forbury, and under the shadow of the church of St.
Laurence, which will be easily recognized by many who remember
how it retained its character as a settlement for prim old ladies, of the
kind described by Miss Mitford, until within quite recent times.
“Of the public amusements of the town, as I remember it at bonny
fifteen,” she continues, “these were sober enough. Ten years before,
clubs had flourished; and the heads of houses had met once a week
at the King’s Arms for the purpose of whist-playing; whilst the ladies,
thus deserted by their liege lords, had established a meeting at each
other’s mansions on club-nights, from which, by way of retaliation,
the whole male sex was banished,” save one. “At the time, however,
of which I speak, these clubs had passed away; and the public
diversions were limited to an annual visit from a respectable
company of actors, the theatre being, as is usual in country places,
very well conducted and exceedingly ill attended; to biennial
concerts, equally good in their kind, and rather better patronised; and
to almost weekly incursions from itinerant lecturers on all the arts
and sciences, and from prodigies of every kind, whether three-year-
old fiddlers or learned dogs. There were also balls in their spacious
and commodious townhall, which seemed as much built for the
purposes of dancing as that of trying criminals. Public balls there
were in abundance; but at the time of which I speak they were of
less advantage to the good town of Belford than any one, looking at
the number of good houses and of pretty young women, could well
have thought possible.”
These few extracts—space forbids a larger selection—are
sufficient, we think, to prove how keen was the observing eye and
how critical was the mind of Miss Mitford at this time when,—to use
her own phrase—“I was a very young girl and, what is more to the
purpose, a very shy one, so that I mixed in none of the gaieties,” a
statement which seems to lend support to the current saying that
“the onlooker sees most of the game.”
So far we have dealt with the Reading life as dating from 1797, but
it is important to note that Miss Mitford speaks of a short residence in
the town when she was but four years of age, and this would give us
the year 1791. Unfortunately no proof for or against this is available,
so far as we know, and we should scarcely have thought it worthy of
mention but for another statement which she makes in her
Recollections the authenticity of which it would be well to at least,
attempt to clear up.
The statement has reference to the interest which Samuel Taylor
Coleridge evinced in one of her earliest literary efforts, Christina; or,
The Maid of the South Seas, when it was being prepared for press at
about the year 1811. This interest she ascribes to “Mr. Coleridge’s
kind recognition of my father’s exertions” in his behalf and relates to
that romantic period of the poet’s life when, in the December of
1793, he suddenly enlisted as a common soldier in the 15th Light
Dragoons under the nom de guerre of “Silas Tomkyn Cumberbatch.”
We have it on good authority that on December 4, 1793, he was
sent, with other raw recruits, to be drilled with his regiment, then
garrisoned at Reading, from which date until his discharge on April
10, 1794, he clearly proved his unfitness for the calling of a man-at-
arms.
The story of his discharge has been variously related, but all are
agreed that his identity was revealed by his being overheard by
certain of his officers reciting Greek lines, to say nothing of the polish
which, scholar as he was, he could not disguise. The circumstance
was sufficiently unique in those days—the gentleman ranker was a
growth of later date—to occasion inquiries, and these resulted in
communications with his friends, who came, identified, and bought
him out. One of these officers was Captain Ogle, eldest son of that
Dean of Winchester to whom, as we have noted in the earliest
chapter of this book, Dr. Mitford was on the visit which resulted in his
introduction to Miss Russell.
Miss Mitford’s account of the event is somewhat circumstantial, for
she relates that as Dr. Ogle was on a short visit to the Mitfords, the
opportunity of calling upon his father was gladly embraced by the
son, who, in the course of conversation, recorded the unusual
incident of the learned recruit, with the result that “one of the
servants waiting at table” was “induced to enlist in his place,” and the
“arrangement for his [Coleridge’s] discharge took place at my
father’s house at Reading.”
The dates relative to Coleridge’s enlistment and discharge are
incontrovertible, therefore in view of the lack of evidence to support
the idea of the Mitfords being in Reading in 1794, we are inclined to
doubt—as others have doubted—the authenticity of Miss Mitford’s
narrative, suggesting rather, that having heard this romantic story,
many years after—possibly from the lips of Captain Ogle himself—
she readily assumed, with the licence of literary folk in general, that
the incident took place as she recorded it.
FOOTNOTES:
[8] He represented Reading in Parliament, 1802-1820.
CHAPTER VI
BERTRAM HOUSE
Bertram House was at last finished and the beginning of 1806 saw
the Mitfords in residence. In the matter of furnishing the Doctor had
spared no expense, everything being new and of the latest pattern,
in fact the best that a fashionable London upholsterer could supply.
Of the pictures we know that the walls were well covered and that
the collection included a Gainsborough, a pair of female heads by
Greuze, and a portrait of the Doctor by Opie. We have already seen,
in Mrs. Mitford’s description of the stone-laying ceremony, that they
were attended by “the two men-servants on horseback”; this hints at
a fairly complete retinue having been installed at the Reading house,
but it was considerably augmented when the arrangements were
completed at Grazeley. Appearances counted for much in the district
and the Doctor was not the man to let slip such a grand opportunity
for ostentatious display.
His hospitality was profuse and indiscriminate, resulting in a
house-warming which extended over quite a lengthy period. As an
incentive—had he need of one—Dr. Mitford had recently been
appointed as one of the County magistrates, a tribute of appreciation
from the Whigs, of whose cause he was an earnest partizan, which
gave him an immediate rise in social status.
In time, of course, the family settled down to a more or less
ordered form of life, so ordered indeed that the Doctor created as
many excuses as possible to cover his frequent journeys to Town
and his clubs. There was sport in plenty to be had in the
neighbourhood, and of this the Doctor took full advantage, being a
familiar figure around the countryside with his gun and spaniels.
Then, too, there were the coursing meetings—the famous meetings
at Ilsley and private matches arranged between friends—none of
which were considered complete unless the Doctor were present or
his famous kennel represented. Throughout Miss Mitford’s letters,
occasionally to her father and often to friends, there are frequent
references to the greyhounds whose names, in accordance with a
custom prevalent then and still fashionable, all began with the letter
M in token of their ownership. Thus, to name only a few, we have
Mia, Manx, Marmion (a notable dog this, with an equally notable son
of the same name), Mogul, Miller, Moss-Trooper, and Mopy. For all of
these Miss Mitford ever exhibited the greatest affection, and in those
cases where a spaniel grew too old to follow the gun or a greyhound
too stiff to be matched, an asylum under Miss Mitford’s immediate
eye and care was immediately provided, and the creature was
henceforth looked upon as her own.
Taking advantage of this motherliness to dumb animals her father
frequently handed over to her some specially valuable dog from
whose later exploits as a courser he expected much. Apparently,
however, the real reason for the supposed gift was not disclosed,
with the result that when the dog was eventually removed the little
mistress gave vent to her annoyance in no measured tones.
“It is a most extraordinary thing,” she says in one communication
to her father, “that I never can have a dog that I like but you
immediately take it from me and burthen me with the care of some
detestable brute whom you in your eternal caprice fancy a good one.
Observe, however, that in giving up my own darling Mordor, I bargain
that that sulky, ungrateful, mangy beast Marmion shall be sent off as
soon as you come home, and that I shall again have my sweet
Marian to pet and comfort me.”
This was not, of course, a serious outburst, but merely the
explosion of what she doubtless considered a truly righteous
indignation, for, although she was no sportswoman, her love for her
father gave her an interest in his pursuits, and she shared with him
to the full the joy of triumph and the sorrow of defeat, while to
disparage the Mitford kennel was to offer her a personal affront. On
the other hand, she was quick to convey to the Doctor any item of
praise which she overheard or might have addressed to her. “We
called yesterday at the Fawcetts’, and the old General said he had
kept greyhounds and seen many thousands, but had never had an
idea of perfect and consummate beauty until he saw her” [a
reference, in a letter to her father, to Mia, one of the hounds].
She had a strong dislike to equestrian exercise—the rides of
babyhood across the Alresford downs with her father could not count
in this connexion—and although every inducement was offered her
to ride, an inglorious fall from a donkey quickly settled her
convictions as to her horsemanship, and her one and only riding-
habit was forthwith converted into a winter-gown. Strictly speaking,
the greater portion of her time was spent at home with her mother,
receiving visitors or lying for hours at a time on the sofa, where she
would devour a great quantity of books at a pace which, having
regard to the extraordinary knowledge she imbibed from her reading,
was truly astonishing.[9] At other times the little green chariot, their
favourite equipage, would be ordered out, calls would be returned
and the drive be possibly extended to Reading, where there would
always be plenty of shopping to do and calls to be made on the old
neighbours and friends who would have the latest news from Town
or the latest gossip of their immediate circle to retail.
With a desire to augment his income, which must have been
seriously depleted by the building operations and by the subsequent
reckless expenditure on the household, the Doctor now began to
indulge in a series of hazardous enterprises, which, with all a
gambler’s insistence, he pursued intently the while they dragged him
deeper and deeper into the mire. One of these was an extensive
speculation in coal in which he engaged with a brother of M. St.
Quintin. For this he supplied the whole of the capital in expectation of
a return of £1,500 a year, but the whole thing was a failure and, with
the exception of about £300, the capital was lost. Another
Frenchman, a man of ingenious ideas but no money wherewith to
put them to practical use, found a ready supporter in the Doctor, who
was induced to advance £5,000 on the strength of a paper scheme.
This man was the Marquis J. M. F. B. de Chabannes, and his
scheme, a supposed improved method for the lighting and heating of
houses, was embodied in a booklet which he published in 1803 with
the comprehensive title of Prospectus d’un Projet pour la
Construction de Nouvelles Maisons, Dont tous les calculs de détails
procureront une très-grande Economie, et beaucoup de
Jouissances. Unfortunately for its promoters, the scheme did not
catch on with the public, the Marquis returned to France and the
deluded Doctor continued for years to spend good money in the
hope of recovering that which was irrevocably lost by suing the
Marquis in the French courts, efforts which were all vain.
Meanwhile his fever for gambling grew apace and his absences
from home were more and more frequent and prolonged, and the
two women, being left much to themselves, conceived the notion of
arranging and copying out for the press a collection of verses
composed by the reverend father of Mrs. Mitford, Dr. Russell. They
took considerable pains with this, to which was added a special
preface by Mrs. Mitford, and when the packet was ready it was
forwarded to Dr. Mitford, in Town, with a request that he should find a
publisher and get as much as he could for it. Unfortunately, the
sanguine editors were disappointed, for no publisher sufficiently
enterprising could be found to accept the manuscript, although
sundry extracts did subsequently find a certain publicity within the
pages of the Poetical Register.
Following closely upon this effort, and in the May of 1806, Miss
Mitford went for a few days on a visit to London as the guest of
Monsieur and Madame St. Quintin, her old schoolmaster and his
wife. A short round of festivities had been arranged for her benefit,
including a visit to the Exhibition of Water Colours, evenings at the
theatres and, what appears to have been a great treat for the
impressionable Miss, some hours of two days which were spent at
Westminster Hall looking on at the trial of Lord Melville[10] and
listening to the speeches, and for which the Doctor, then in Town and
staying at Richardson’s Hotel in Covent Garden, had procured
tickets. She had now been absent from London for over three years
and, no doubt, extracted a great deal of pleasure from her visit and
its reunion with Fanny Rowden and Victoire St. Quintin, M. St.
Quintin’s sister, with both of whom, together with the Doctor, the
round of sight-seeing was enjoyed.
Mrs. Mitford stayed at home, but was kept well-posted in all the
news by the inevitable letters, full of critical details, from her dutiful
daughter. From one of these, dated from Hans Place, May 12, 1806,
we quote:—
“I have much to tell you, but it can scarcely be compressed within
the bounds of a letter. On Thursday, after I wrote, Miss Ayrton, Miss
Carp, papa, and I went to the Exhibition. There are some
uncommonly fine pictures, and it is even better worth seeing than
last year. In the evening, Victoire, Miss A. and myself went with papa
to the play to see The Provoked Husband and The Forty Thieves.
Miss Duncan in Lady Townley is most admirable. I do not much
admire Elliston as her husband. The Forty Thieves is a very
magnificent spectacle, but nothing more; for the language and music
are equally vulgar and commonplace. On Friday morning we went to
Oxford Street. I was extravagant enough to give half a guinea for a
dress skirt for myself, which I wore the next day to the trial. We were
rather disappointed in Mr. Romilly.[11] The speech in itself was
beautiful beyond description; but he wants animation, and drops his
voice at the end of every sentence.... Miss Rowden, papa, and I are
going to see Henry the Eighth to-night, and we are going to
Westminster Hall to-morrow.... I shall hope to return Thursday or
Friday; for, though I am greatly amused here, I am never quite happy
without my dear, dear mother.”
Two days later this was followed by a still more characteristic
effusion. The second day at Westminster Hall decided her that: “Mr.
Romilly is charming and interesting; but my first and greatest
favourite is Mr. Whitbread. Mr. Plumer is rather an inelegant speaker,
though very animated. I have promised papa to write some verses to
Mr. Whitbread. He has even superseded Mr. Fox in my good graces.
I did not tell you, I believe, that I had the happiness of seeing Mr. Fox
mount his horse on Saturday. I shall never again contend for his
beauty. He was obliged to lean on two people, and looked so sallow
and pale in the face, and so unwieldy in person, that I am obliged to
yield our long-disputed point.” Rather hard on poor Mr. Fox, whom,
hitherto, this exuberant young person had worshipped as a hero,
even to the extent of removing her watch-stand from the head of her
bed that it might give place to a bust of this gentleman which the
Doctor had sent from Town. On this occasion it was a case of “Off
with the old love and on with the new” in double-quick time, for,
continuing, she says: “To make me amends my new favourite is what
even you would call exquisitely handsome; a most elegant figure,
and a voice which I could listen to with transport, even if he spoke in
an unknown language. Mr. Plumer attacked him with the most
virulent irony and ridicule; and Mr. W. stood with his face turned
towards him and leant upon the desk, smiling the whole time, with
the most fascinating good humour. You know I am always an
enthusiast; but at present it is impossible to describe the admiration I
feel for this exalted character.”
We quote these extracts with no thought of ridiculing the ardent
partisan, but as a fore-shadowing of that enthusiasm and that quick
impressibility which ever seemed to dominate Miss Mitford’s life;
characteristics which often led her into excesses of transport at the
discovery, or supposed discovery, of some noble trait in the
characters of those who came within her ken, only to be as quickly
repented of; often giving unintentional pain to others and resulting in
an infinitude of trouble and annoyance to herself. Despite this
temperamental defect, however, and while her friends looked on
amazed at her infidelity, there was one to whom she remained
unwaveringly faithful to the end, though this object of her great
affection was the least worthy of all who came into her life.
Mr. Whitbread, favoured man, was the immediate recipient of

some verses from his ardent admirer. They reached him, ten days
after his Westminster display of elegance and fortitude, through Dr.
Mitford, to whom they were posted from Bertram House under cover
of the following ingenuous letter: “May 24, 1806.—I claim great merit,
my dear darling, in sending you the enclosed lines, for I am not
satisfied with them; but I would sooner mortify my own vanity by
sending you bad verses, than break my promise by withholding
them. I have called them impromptu to excuse their incorrectness;
and though some may suspect them to be an impromptu fait à loisir,
you must not betray the secret. From a perfect consciousness of my
own enthusiasm, I have been so much afraid of saying too much,
that I have fallen into the opposite fault and said too little. However, I
had rather be thought anything but a flatterer, though it be in my
opinion impossible to flatter Mr. Whitbread; for what language can
equal his merits? Do not impute the faults and deficiencies in these
lines to my laziness; for I assure you they cost me an infinite deal of
trouble; but they are not good enough to show, and I had rather you
would return them to me immediately. At all events, let me know how
you like them, and what you have done with them.”
Not to be misled by the feigned artlessness of his daughter’s
concluding sentences, the Doctor, as we have said, passed on the
verses to Mr. Whitbread, who was pleased to acknowledge and
eulogise them; and since they deserve it we give them below:—
Impromptu on Hearing Mr. Whitbread Declare in
Westminster Hall, on Friday, May 16, 1806, that He
“Fondly Trusted His Name would Descend with Honour
to Posterity.”
The hope of Fame thy noble bosom fires,

Nor vain the hope thy ardent mind inspires;
In British breasts, whilst Purity remains,
Whilst Liberty her blest abode retains,
Still shall the muse of History proclaim
To future ages thy immortal name.
And while fair Scotia weeps her favoured son,
By place corrupted and by power undone,
England with pride her upright patriot sees,
And Glory’s brightest wreath to him decrees.
FOOTNOTES:
[9] A list kept as a check on the Circulating Library account for the
years 1806 to 1811 inclusive, is a sufficient indication of this, the
number for one month alone totalling fifty-five volumes and
ranging through Fiction, Belles-Lettres, Travel and Biography.
[10] Impeached for malversation in his office as Treasurer of the
Navy. The trial lasted sixteen days. Whitbread led for the
Impeachers; Plumer—afterwards Master of the Rolls—ably
defended and secured his acquittal.
[11] Sir Samuel Romilly, Solicitor-General.
CHAPTER VII
THE TRIP TO NORTHUMBERLAND
With a proper and increasing pride in his clever daughter, the Doctor
now conceived the idea of taking her with him on an extended trip
into Northumberland, thereby affording her some acquaintance with
the scenes amidst which his family had lived for generations, a trip
which would serve the double purpose of impressing the girl with a
sense of the importance of her ancestors and present relations and
of introducing her to the latter who, although they must have heard of
her, had never yet seen her.
The journey was begun on Saturday, September 20, 1806, the first
stage, that from Reading to London being by coach. From London
they travelled in the carriage of Nathaniel Ogle, who personally
conducted them to his own place in Northumberland, from which
they were to make their various excursions in the district. Mrs.
Mitford did not accompany them, but was kept well informed, as
usual, by her daughter.
The first letter is dated from Royston, September 21: “We had a
very long interval between the parting from my most beloved darling
and the leaving Reading. The coach was completely full; and it was
fortunate papa had secured a place on the box, where he continued
during the whole journey. The company in the inside had the merit of
being tolerably quiet; and I do not remember any conversation which
lasted longer than a minute. I, certainly, ought not to complain of
their silence, as I was more than equally taciturn, and scarcely spoke
during the whole way. I was quite low-spirited, but never less
fatigued by travelling. Both Mr. Joy and Dr. Valpy[12] met us before we
left Reading, and M. St. Quintin and Victoire met us at the Bath
Hotel. As soon as Victoire left me, I retired to bed, under the idea of
pursuing our journey early in the morning. It was, however, half-past
ten before Mr. Ogle got up, and we did not leave town till twelve. We
employed the interval in going to the bookseller’s for a Cobbett, and
bought a Cary’s Itinerary, an edition of Peter Pindar, and a few plays.
The Edition of P. P. which we bought cheap, remains in town; but the
others are our travelling companions. We went by Enfield to see
Mary Ogle, and finding them at dinner we dined at Mrs. Cameron’s;
we then changed horses at Waltham Cross; again at Wade’s Mill;
and are just arrived here, where we sleep to-night. Mr. Ogle is
extremely pleasant, and the carriage very convenient. We went the
two first stages on the box of the barouche. I need not tell you, my
dearest darling, that we felt nothing so much as the loss of your
society; and I have wished myself at home fifty times in the last
twenty-four hours, to be again with my dear mamma.”
Apart from the interest which, in these days, is always attached to
an old-time account of stage-travel the letter is interesting by reason
of the variety of literature purchased for perusal on the journey. The
Cobbett referred to would probably be Cobbett’s Political Register
(then being issued in parts), and intended for the Doctor’s personal
reading; he being not only an admirer but an intimate friend of the
outspoken reformer. Cary’s Itinerary was, of course, the well-known
road-book and constant companion of all who travelled in stage-
coach days; though why Miss Mitford was not content with her
dainty, green-leather-covered copy of Bowles’ Post-Chaise
Companion in two vols.—now a valued possession of the author’s—
is difficult to understand, unless it was overlooked in the hurry and
excitement of departure. Peter Pindar’s Works, then just completed
in five vols., would be a valuable addition to the library at home, but
the purchase of the plays is significant, proving the influence which
Fanny Rowden had exercised on the mind of her pupil, inculcating a
taste for the Drama which was to be of lasting importance.
The next letter is written from Little Harle Tower—a small place
about fourteen miles from Morpeth—and is dated Sunday Evening,
September 28.—
“I arrived here with Lady Charles,[13] about two hours since, my
dearest mamma; and I find from papa that in his letter to you to-night
he never mentioned that the irregularity of the post, which never
goes oftener than three times a week from hence, will prevent our
writing again till Wednesday, when we go to Sir William Lorraine’s,
and hope to get a frank from Colonel Beaumont, whom we are to
meet there. It is only by Lord Charles going unexpectedly to Morpeth
that I am able to write this, merely to beg you not to be alarmed at
not hearing oftener. I imagine papa has told you all our plans, which
are extremely pleasant. Lord and Lady Charles stay longer in the
country on purpose to receive us, and have put off their visit to
Alnwick Castle that they may take me there, as well as to Lord
Grey’s, Colonel Beaumont’s, and half a dozen other places.”
The reference in this letter to a “frank” is one which frequently
occurs in Miss Mitford’s correspondence. It was, as Sir Rowland Hill
once said, an “expedient for saving postage”—“discreditable shifts”
another writer called them. In the days before the institution of Penny
Postage—an event which put an end to “franking”—Members
of Parliament enjoyed the privilege of having their letters delivered
and despatched free of charge. To secure this, members had merely
to write their names on the covers to ensure free passage through
the post, and frequently furnished their friends with packets of franks
which were placed aside for use as occasion required. This latter
expedient was, of course, a flagrant abuse of the privilege, and in
one year it was computed that, had postage been paid on the
franked correspondence, the revenue would have been increased by
£170,000! In an endeavour to check this abuse it was enacted that
the whole superscription must be in the handwriting of the Member,
and that the frank was only available on the date (which it was
necessary to name) which was on the cover. While the regulation
certainly diminished the quantity of franking it did not put an end to
the use of the privilege by other than Members, to whom it became
the custom to despatch an accumulated batch of letters, intended for
a number of people, with explicit instructions as to their destinations.
The annoyance caused to Members, and the general confusion
which sometimes resulted from this practice, may be better imagined
than described. Miss Mitford herself gives us an amusing account of
the troubles and trials of those who both used and abused the
franking privilege, in her sketch on “The Absent Member,” in Belford
Regis.
In the next letter which Miss Mitford wrote we have a record of
some amusing table-talk, essentially feminine in character and
which, undoubtedly, greatly impressed the observant young person
who overheard it. It is addressed still from Little Harle Tower, dated
October 3, and after a short description of the scenery, and the mud
—which caused her to beg to be excused from such excursions in
the future—she relates an account of a dinner at Sir William
Lorraine’s at which Colonel and Mrs. Beaumont were of the party.—
“Mrs. B. was so polite as to express great regret that, as she was
going from home, she could not see us at her house, but hoped,
when next we came to Northumberland, we should come to see
them at Hexham Abbey. She is a very sweet woman.... Mrs. B. told
Lady Charles that they received last year a hundred thousand
pounds from their lead mines in Yorkshire; and they never make less
than eighty thousand, independent of immense incomes from their
other estates. Mrs. B. was dressed in a lavender-coloured satin, with
Mechlin lace, long sleeves, and a most beautiful Mechlin veil. The
necklace she wore was purchased by her eldest son, a boy of
eleven, who sent it from the jeweller’s without asking the price. It is
of most beautiful amethysts; the three middle stones are an inch and
a half long and an inch wide; the price was nine hundred guineas.
Mrs. B. wished to return it; but the Colonel not only confirmed the
purchase, but gave his son some thousands to complete the set of
amethysts by a bandeau and tiara, a cestus for the waist, armlets,
bracelets, brooches, sleeve-clasps, and shoe-knots. All these she
wore, and I must confess, for a small dinner-party appeared rather
too gaily decorated, particularly as Lady Lorraine’s dress was quite
in the contrary extreme. I never saw so strong a contrast ... Colonel
Beaumont is generally supposed to be extremely weak, but I sat next
him at dinner, and he conducted himself with infinite propriety and
great attention and politeness; yet when away from Mrs. Beaumont,
he is (they say) quite foolish, and owes everything to her influence.”
Added to this cryptic description—cryptic because, read it how we
will, we cannot be sure that there is not a subtle touch of sarcasm in
the words—is a shrewd observation on another visitor whom she
calls Mr. M.
“I told you I was not enamoured of Mr. M., and I will now describe
him to you.... He is an oddity from affectation; and, I often think, no
young man affects singularity when he can distinguish himself by
anything better. He affects to despise women, yet treats them with
great respect; and he makes the most extraordinary exertions to
provoke an argument, from which he generally escapes by some
whimsical phrase.”
The letter concludes with a long list of festivities which are to take
place in honour of her visit. Following on these, they journeyed to
Kirkley, Mr. Ogle’s seat, whither it was originally intended they should
travel direct but were deterred from so doing by the hospitality
offered en route. As a matter of fact their stay at Kirkley was a short
one, due to the same cause which had prevented their earlier arrival.
The only letter addressed from Kirkley is dated Wednesday
morning, October 8:—
“We arrived here on Monday at about three o’clock; received with
great glee by the Squire, and, after taking a short walk in the garden,
returned to dress. We had some time to wait for Lord and Lady
Charles, who did not arrive before half-past five or near six, and
even then undressed. They had been detained by the axle-tree
breaking down, and the detestable roads. Without their waiting to
dress, we immediately sat down to dinner and spent a most
delightful day. In the evening we found a manuscript play which had
been sent last year for Mr. Sheridan’s perusal.[14] It is taken from a
very striking story in the Canterbury Tales, of which I have forgotten
the title.... I read it aloud to the ladies, and the gentlemen played
billiards, and occasionally visited us. The play, which bears the name
of ‘Sigendorf,’ is really extremely interesting, and much better, as to
language, than most modern productions. Sheridan had never
looked at it, and Mr. Ogle lent it to Lady Charles.
“Yesterday morning, after a long walk, Lord and Lady C. left us.
We had an excellent dinner, and amused ourselves in the evening
with the ‘Liber Veritatis,’ which is, as you may remember, a very
expensive collection of two hundred of Claude Lorraine’s sketches,
published by Boydell.
“We are going in about an hour to Little Harle ... for Mr. Ogle and
papa remain here together. We go to-morrow to Alnwick and return
the same night. To-morrow is expected to be a very full day at the
Castle, on account of the Sessions Ball. The ladies—the married
ones I mean—go in Court dresses, without hoops, and display their
diamonds and finery on the occasion.
“Mr. Ogle is quite a man of gallantry and makes his house
extremely pleasant. We talk of coming to see him again next week,
when my cousin Mary and I are left to keep house alone at Morpeth,
and my uncle and aunt go to Little Harle Tower.”
From Morpeth, on October 11, was despatched a very long letter,
too long indeed for quotation in full, but from which we must give a
few extracts. It begins:—
“In papa’s letter of yesterday, my dearest darling mamma, he
promised that I would write you a long one to-day, and I certainly
owe you one in return for the very entertaining epistle I received
yesterday. After we left Kirkley, we called at Belsay, and saw Lady
Monck and the little Atticus, who was born at Athens fifteen months
since. He is a very fine boy, very like Sir Charles. Belsay is a very old
castle, and its eccentric possessor has done all he possibly could
desire to render it still more outré by stopping up the proper road,
and obliging us to approach this fine specimen of Gothic architecture
through the farm-yard. We arrived at Little Harle to dinner; and you
would have been greatly amused at my having my hair cut by Lord
Charles’s friseur, who is by occupation a joiner, and actually
attended with an apron covered with glue, and a rule in his hand
instead of scissors. He, however, performed his office so much to my
satisfaction, that I appointed him to dress my hair the next morning
for my visit to Alnwick. While I was thus employed, Lady Swinburne
called on purpose to see me. Lady Charles said I was out walking.
She is, you know, niece to the Duke of Northumberland, and I
regretted not seeing her.

PDF Advances in Computer Vision Proceedings of The 2019 Computer Vision Conference CVC Volume 1 Kohei Arai Ebook Full Chapter

Uploaded by

Copyright:

Available Formats

You might also like

PDF Advances in Computer Vision Proceedings of The 2019 Computer Vision Conference CVC Volume 1 Kohei Arai Ebook Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PDF Advances in Computer Vision Proceedings of The 2019 Computer Vision Conference CVC Volume 1 Kohei Arai Ebook Full Chapter

Uploaded by

Copyright:

Available Formats

Advances in Computer Vision:

Proceedings of the 2019 Computer

Proceedings of the Future Technologies Conference FTC

Proceedings of the Future Technologies Conference (FTC)

Intelligent Computing Proceedings of the 2020 Computing

Applied Computer Vision and Image Processing

Proceedings of the Future Technologies Conference FTC

Intelligent Computing Proceedings of the 2020 Computing

Intelligent Computing Proceedings of the 2020 Computing

Computer Vision Systems 12th International Conference

** Indexing: The books of this series are submitted to ISI Proceedings,

More information about this series at http://www.springer.com/series/11156

ISSN 2194-5357 ISSN 2194-5365 (electronic)

Deep Learning for Detection of Railway Signs and Signals . . . . . . . . . . 1

ISRGAN: Improved Super-Resolution Using Generative

A Probabilistic Superpixel-Based Method for Road Crack

Standardization of the Shape of Ground Control Point (GCP)

Quantitative Comparison of White Matter Segmentation

Decision Making with Linguistic Information for the Development

Georgios Karagiannis1,2(B) , Søren Olsen1 , and Kim Pedersen1

Abstract. Major railway lines need advance management systems

Keywords: Railway · Object detection · Object recognition ·

networks. However, with currently available methods, mapping of railway assets

2.1 Previous Approaches

The research on automatic object detection along railways is sparse, compared

2.2 Convolutional Neural Networks

Figure 3 shows examples of the same class of signals (Signal1). It is important

Another important aspect of the samples is their size. Figure 6 is a histogram

3.2 Railway vs Road Signs and Signals

Here, it is important to highlight the diﬀerence between the problem described in

signals needs to be able to distinguish objects based on ﬁne details. Finally, in

4.1 Faster R-CNN

bounding box compared to the ground-truth bounding boxes. To reduce redun-

4.2 Evaluation Method

Figure 11 is a confusion matrix that summarises the performance of the

Zhangsihao Yang, Haoliang Jiang, and Lan Zou(&)

Carnegie Mellon University, Pittsburgh, PA 15213, USA

Abstract. This article proposes a data-driven methodology to achieve fast

Keywords: Design support Data analysis 3D representation

Three-dimensional (3D) representations of real-life objects have been an essential

© Springer Nature Switzerland AG 2020

Mr. Whitbread, favoured man, was the immediate recipient of

The hope of Fame thy noble bosom fires,

THE TRIP TO NORTHUMBERLAND

You might also like