Intelligent Computing Theories and

Application De-Shuang Huang

Visit to download the full and correct content document:
https://textbookfull.com/product/intelligent-computing-theories-and-application-de-shu
ang-huang/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Intelligent Computing Theories and Application 16th

International Conference ICIC 2020 Bari Italy October 2
5 2020 Proceedings Part I De-Shuang Huang

https://textbookfull.com/product/intelligent-computing-theories-
and-application-16th-international-conference-icic-2020-bari-
italy-october-2-5-2020-proceedings-part-i-de-shuang-huang/

Intelligent Computing Theories and Application 13th

International Conference ICIC 2017 Liverpool UK August
7 10 2017 Proceedings Part I 1st Edition De-Shuang
Huang
https://textbookfull.com/product/intelligent-computing-theories-
and-application-13th-international-conference-
icic-2017-liverpool-uk-august-7-10-2017-proceedings-part-i-1st-
edition-de-shuang-huang/

Intelligent Computing Theory 10th International

Conference ICIC 2014 Taiyuan China August 3 6 2014
Proceedings 1st Edition De-Shuang Huang

https://textbookfull.com/product/intelligent-computing-
theory-10th-international-conference-icic-2014-taiyuan-china-
august-3-6-2014-proceedings-1st-edition-de-shuang-huang/

Intelligent Computing Methodologies 10th International

Conference ICIC 2014 Taiyuan China August 3 6 2014
Proceedings 1st Edition De-Shuang Huang

https://textbookfull.com/product/intelligent-computing-
methodologies-10th-international-conference-icic-2014-taiyuan-
china-august-3-6-2014-proceedings-1st-edition-de-shuang-huang/
Intelligent Computing Methodologies 16th International
Conference ICIC 2020 Bari Italy October 2 5 2020
Proceedings Part III De-Shuang Huang

https://textbookfull.com/product/intelligent-computing-
methodologies-16th-international-conference-icic-2020-bari-italy-
october-2-5-2020-proceedings-part-iii-de-shuang-huang/

Intelligent Computing in Bioinformatics 10th

International Conference ICIC 2014 Taiyuan China August
3 6 2014 Proceedings 1st Edition De-Shuang Huang

https://textbookfull.com/product/intelligent-computing-in-
bioinformatics-10th-international-conference-icic-2014-taiyuan-
china-august-3-6-2014-proceedings-1st-edition-de-shuang-huang/

Application of Soft Computing and Intelligent Methods

in Geophysics Alireza Hajian

https://textbookfull.com/product/application-of-soft-computing-
and-intelligent-methods-in-geophysics-alireza-hajian/

Soft Computing Theories and Applications Proceedings of

SoCTA 2019 Advances in Intelligent Systems and
Computing Millie Pant (Editor)

https://textbookfull.com/product/soft-computing-theories-and-
applications-proceedings-of-socta-2019-advances-in-intelligent-
systems-and-computing-millie-pant-editor/

Intelligent Fault Diagnosis and Accommodation Control

1st Edition Sunan Huang (Author)

https://textbookfull.com/product/intelligent-fault-diagnosis-and-
accommodation-control-1st-edition-sunan-huang-author/
 De-Shuang Huang
Kang-Hyun Jo
Xiao-Long Zhang (Eds.)
LNCS 10955

Intelligent Computing
Theories and Application
14th International Conference, ICIC 2018
Wuhan, China, August 15–18, 2018
Proceedings, Part II

123
Lecture Notes in Computer Science 10955
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zurich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology Madras, Chennai, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7409
De-Shuang Huang Kang-Hyun Jo
•

Xiao-Long Zhang (Eds.)

Intelligent Computing
Theories and Application
14th International Conference, ICIC 2018
Wuhan, China, August 15–18, 2018
Proceedings, Part II

123
Editors
De-Shuang Huang Xiao-Long Zhang
Tongji University Wuhan University of Science
Shanghai and Technology
China Wuhan City
China
Kang-Hyun Jo
University of Ulsan
Ulsan
Korea (Republic of)

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-319-95932-0 ISBN 978-3-319-95933-7 (eBook)
https://doi.org/10.1007/978-3-319-95933-7

Library of Congress Control Number: 2018947576

LNCS Sublibrary: SL3 – Information Systems and Applications, incl. Internet/Web, and HCI

© Springer International Publishing AG, part of Springer Nature 2018

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, express 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.

Printed on acid-free paper

This Springer imprint is published by the registered company Springer International Publishing AG
part of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

The International Conference on Intelligent Computing (ICIC) was started to provide

an annual forum dedicated to the emerging and challenging topics in artificial intelli-
gence, machine learning, pattern recognition, bioinformatics, and computational biol-
ogy. It aims to bring together researchers and practitioners from both academia and
industry to share ideas, problems, and solutions related to the multifaceted aspects of
intelligent computing.
ICIC 2018, held in Wuhan, China, August 15–18, 2018, constituted the 14th
International Conference on Intelligent Computing. It built upon the success of ICIC
2017, ICIC 2016, ICIC 2015, ICIC 2014, ICIC 2013, ICIC 2012, ICIC 2011, ICIC
2010, ICIC 2009, ICIC 2008, ICIC 2007, ICIC 2006, and ICIC 2005 that were held in
Liverpool, UK, Lanzhou, Fuzhou, Taiyuan, Nanning, Huangshan, Zhengzhou,
Changsha, China, Ulsan, Korea, Shanghai, Qingdao, Kunming, and Hefei, China,
respectively.
This year, the conference concentrated mainly on the theories and methodologies as
well as the emerging applications of intelligent computing. Its aim was to unify the
picture of contemporary intelligent computing techniques as an integral concept that
highlights the trends in advanced computational intelligence and bridges theoretical
research with applications. Therefore, the theme for this conference was “Advanced
Intelligent Computing Technology and Applications.” Papers focused on this theme
were solicited, addressing theories, methodologies, and applications in science and
technology.
ICIC 2018 received 632 submissions from 19 countries and regions. All papers went
through a rigorous peer-review procedure and each paper received at least three
reviews. Based on the review reports, the Program Committee finally selected 275
high-quality papers for presentation at ICIC 2018, included in three volumes of pro-
ceedings published by Springer: two volumes of Lecture Notes in Computer Science
(LNCS) and one volume of Lecture Notes in Artificial Intelligence (LNAI).
This volume of Lecture Notes in Computer Science (LNCS) includes 93 papers.
The organizers of ICIC 2018, including Tongji University, Wuhan University of
Science and Technology, and Wuhan Institute of Technology, made an enormous effort
to ensure the success of the conference. We hereby would like to thank the members
of the Program Committee and the referees for their collective effort in reviewing and
soliciting the papers. We would like to thank Alfred Hofmann, executive editor at
Springer, for his frank and helpful advice and guidance throughout and for his con-
tinuous support in publishing the proceedings. In particular, we would like to thank all
the authors for contributing their papers. Without the high-quality submissions from the
authors, the success of the conference would not have been possible. Finally, we are
VI Preface

especially grateful to the IEEE Computational Intelligence Society, the International

Neural Network Society, and the National Science Foundation of China for their
sponsorship.

May 2018 De-Shuang Huang

Kang-Hyun Jo
Xiao-Long Zhang
 ICIC 2018 Organization

General Co-chairs

De-Shuang Huang, China

Huai-Yu Wu, China
Yanduo Zhang, China

Program Committee Co-chairs

Kang-Hyun Jo, Korea

Xiao-Long Zhang, China
Haihui Wang, China
Abir Hussain, UK

Organizing Committee Co-chairs

Hai-Dong Fu, China

Yuntao Wu, China
Bo Li, China

Award Committee Co-chairs

Juan Carlos Figueroa, Colombia

M. Michael Gromiha, India

Tutorial Chair

Vitoantonio Bevilacqua, Italy

Publication Co-chairs

Kyungsook Han, Korea

Phalguni Gupta, India

Workshop Co-chairs

Valeriya Gribova, Russia

Laurent Heutte, France
Xin Xu, China
VIII ICIC 2018 Organization

Special Session Chair

Ling Wang, China

International Liaison Chair

Prashan Premaratne, Australia

Publicity Co-chairs

Hong Zhang, China

Michal Choras, Poland
Chun-Hou Zheng, China
Jair Cervantes Canales, Mexico

Sponsors and Exhibits Chair

Wenzheng Bao, Tongji University, China

Program Committee

Abir Hussain Gaoxiang Ouyang José Alfredo Costa

Akhil Garg Haiying Ma Juan Carlos
Angelo Ciaramella Han Zhang Figueroa-García
Ben Niu Hao Lin Junfeng Xia
Bin Liu Hongbin Huang Junhui Gao
Bing Wang Honghuang Lin Junqing Li
Bingqiang Liu Hongjie Wu Ka-Chun Wong
Binhua Tang Hongmin Cai Khalid Aamir
Bo Li Hua Tang Kyungsook Han
Chunhou Zheng Huiru Zheng Laurent Heutte
Chunmei Liu Jair Cervantes Le Zhang
Chunyan Qiu Jian Huang Liang Gao
Dah-Jing Jwo Jianbo Fan Lida Zhu
Daowen Qiu Jiang Xie Ling Wang
Dong Wang Jiangning Song Lining Xing
Dunwei Gong Jianhua Xu Lj Gong
Evi Syukur Jiansheng Wu Marzio Pennisi
Fanhuai Shi Jianyang Zeng Michael Gromiha Maria
Fei Han Jiawei Luo Siluvay
Fei Luo Jing-Yan Wang Michal Choras
Fengfeng Zhou Jinwen Ma Ming Li
Francesco Pappalardo Jin-Xing Liu Mohd Helmy Abd Wahab
Gai-Ge Wang Ji-Xiang Du Pei-Chann Chang
 ICIC 2018 Organization IX

Ping Guo Tianyong Hao Xingwen Liu

Prashan Premaratne Valeriya Gribova Xinyi Le
Pu-Feng Du Vasily Aristarkhov Xiwei Liu
Qi Zhao Vitoantonio Bevilacqua Xuesong Wang
Qingfeng Chen Waqas Haider Khan Xuesong Yan
Qinghua Jiang Bangyal Xu-Qing Tang
Quan Zou Wei Chen Yan Wu
Rui Wang Wei Jiang Yan Zhang
Sabri Arik Wei Peng Yi Xiong
Saiful Islam Wei Wei Yong Wang
Seeja K. R. Wei-Chiang Hong Yonggang Lu
Shan Gao Weijia Jia Yongquan Zhou
Shanfeng Zhu Weiwei Kong Yoshinori Kuno
Shih-Hsin Chen Wen Zhang Young B. Park
Shiliang Sun Wenbin Liu Yuan-Nong Ye
Shitong Wang Wen-Sheng Chen Zhan-Li Sun
Shuai Li Hong Wenyin Gong Zhao Liang
Stefano Squartini Xiandong Meng Zhendong Liu
Sungshin Kim Xiaoheng Deng Zhenran Jiang
Surya Prakash Xiaoke Ma Zhenyu Xuan
Takashi Kuremoto Xiaolei Zhu Zhihua Zhang
Tao Zeng Xiaoping Liu
Tarık Veli Mumcu Xinguo Lu

Additional Reviewers

Huijuan Zhu Hongguan Liu Xia Lin

Yizhong Zhou Hai-tao Li Yang Wu
Lixiang Hong Jialing Li Xiaoming Liu
Yuan Wang Kai Qian Jing Lv
Mao Xiaodan Huichao Zhong Lin Weizhong
Ke Zeng Huiyan Jiang Jun Li
Xiongtao Zhang Lei Wang Li Peng
Ning Lai Yuanyuan Wang Hongfei Bao
Shan Gao Biao Zhang Zhaoqiang Chen
Jia Liu Ta Zhou Ru Yang
Ye Tang Wei Liao Jiayao Wu
Weiwei Cai Bin Qin Dadong Dai
Yan Zhang Jiazhou Chen Guangdi Liu
Zhang Yuanpeng Mengze Du Jiajia Miao
Han Zhu Sheng Ding Xiuhong Yang
Wei Jiang Dongliang Qin Xiwen Cai
Hong Peng Syed Sadaf Ali Fan Li
Wenyan Wang Zheng Chenc Aysel Ersoy Yilmaz
Xiaodan Deng Shang Xiang Agata Giełczyk
X ICIC 2018 Organization

Akila Ranjith Ling-Yun Dai Geethan Mendiz

Xiao Yang Dongbo Bu Guanghui Li
Cheng Liang Deming Lei Giacomo Donato
Alessio Ferone Deepak Ranjan Nayak Cascarano
José Alfredo Costa Dong Han Giorgio Valle
Ambuj Srivastava Xiaojun Ding Giovanni Dimauro
Mohamed Abdel-Basset Domenico Buongiorno Giulia Russo
Angelo Ciaramella Haizhou Wu Linting Guan
Anthony Chefles Pingjian Ding Ping Gong
Antonino Staiano Dongqing Wei Yanhui Gu
Antonio Brunetti Yonghao Du Gunjan Singh
Antonio Maratea Yi Yao Guohua Wu
Antony Lam Ekram Khan Guohui Zhang
Alfredo Pulvirenti Miao Jiajia Guo-sheng Hao
Areesha Anjum Ziqing Liu Surendra M. Gupta
Athar Ali Moinuddin Sergio Santos Sandesh Gupta
Mohd Ayyub Khan Tomasz Andrysiak Gang Wang
Alfonso Zarco Fengyi Song Hafizul Fahri Hanafi
Azis Ciayadi Xiaomeng Fang Haiming Tang
Brendan Halloran Farzana Bibi Fei Han
Bin Qian Fatih Adıgüzel Hao Ge
Wenbin Song Fang-Xiang Wu Kai Zhao
Benjamin J. Lang Dongyi Fan Hangbin Wu
Bo Liu Chunmei Feng Hui Ding
Bin Liu Fengfeng Zhou Kan He
Bin Xin Pengmian Feng Bifang He
Guanya Cai Feng Wang Xin He
Casey P. Shannon Feng Ye Huajuan Huang
Chao Dai Farid Garcia-Lamont Jian Huang
Chaowang Lan Frank Shi Hao Lin
Chaoyang Zhang Chien-Yuan Lai Ling Han
Zhang Chuanchao Francesco Fontanella Qiu Xiao
Jair Cervantes Lei Shi Yefeng Li
Bo Chen Francesca Nardone Hongjie Wu
Yueshan Cheng Francesco Camastra Hongjun Bai
Chen He Francesco Pappalardo Hongtao Lei
Zhen Chen Dongjie Fu Haitao Zhang
Chen Zhang Fuhai Li Huakang Li
Li Cao Hisato Fukuda Jixia Huang
Claudio Loconsole Fuyi Li Pu Huang
Cláudio R. M. Silva Gai-Ge Wang Sheng-Jun Huang
Chunmei Liu Bo Gao Hailin Hu
Yan Jiang Fei Gao Xuan Huo
Claus Scholz Hongyun Gao Wan Hussain Wan Ishak
Yi Chen Jianzhao Gao Haiying Wang
Dhiya AL-Jumeily Gaoyuan Liang Il-Hwan Kim
 ICIC 2018 Organization XI

Kamlesh Tiwari Liangfang Zhao Muhammad Asghar

M. Ikram Ullah Lali Jing Liang Nadeem
Ilaria Bortone Bo Li Nadir Subaşi
H. M. Imran Feng Li Nagarajan Raju
Ingemar Bengtsson Jianqiang Li Davide Nardone
Izharuddin Izharuddin Lijun Quan Nathan R. Cannon
Jackson Gomes Junqing Li Nicole Yunger Halpern
Wu Zhang Min Li Ning Bao
Jiansheng Wu Liming Xie Akio Nakamura
Yu Hu Ping Li Zhichao Shi
Jaya sudha Qingyang Li Ruxin Zhao
Jianbo Fan Lisbeth Rodríguez Mohd Norzali Hj Mohd
Jiancheng Zhong Shaohua Li Nor Surayahani Suriani
Enda Jiang Shiyong Liu Wataru Ohyama
Jianfeng Pei Yang Li Kazunori Onoguchi
Jiao Zhang Yixin Li Aijia Ouyang
Jie An Zhe Li Paul Ross McWhirter
Jieyi Zhao Zepeng Li Jie Pan
Jie Zhang Lulu Zuo Binbin Pan
Jin Lu Fei Luo Pengfei Cui
Jing Li Panpan Lu Pu-Feng Du
Jingyu Hou Liangxu Liu Iyyakutti Iyappan
Joe Song Weizhong Lu Ganapathi
Jose Sergio Ruiz Xiong Li Piyush Joshi
Jiang Shu Junming Zhang Prashan Premaratne
Juntao Liu Shingo Mabu Peng Gang Sun
Jiawen Lu Yasushi Mae Puneet Gupta
Jinzhi Lei Malik Jahan Khan Qinghua Jiang
Kanoksak Wattanachote Mansi Desai Wangren Qiu
Juanjuan Kang Guoyong Mao Qiuwei Li
Kunikazu Kobayashi Marcial Guerra Shi Qianqian
Takashi Komuro de Medeiros Zhi Xian Liu
Xiangzhen Kong Ma Wubin Raghad AL-Shabandar
Kulandaisamy A. Xiaomin Ma Rafał Kozik
Kunkun Peng Medha Pandey Raffaele Montella
Vivek Kanhangad Meng Ding Woong-Hee Shin
Kang Xu Muhammad Fahad Renjie Tan
Kai Zheng Haiying Ma Rodrigo A. Gutiérrez
Kun Zhan Mingzhang Yang Rozaida Ghazali
Wei Lan Wenwen Min Prabakaran
Laura Yadira Domínguez Mi-Xiao Hou Jue Ruan
Jalili Mengjun Ming Rui Wang
Xiangtao Chen Makoto Motoki Ruoyao Ding
Leandro Pasa Naixia Mu Ryuzo Okada
Erchao Li Marzio Pennisi Kalpana Shankhwar
Guozheng Li Yong Wang Liang Zhao
XII ICIC 2018 Organization

Sajjad Ahmed Vasily Aristarkhov Yongchun Zuo

Sakthivel Ramasamy Vitoantonio Bevilacqua Bei Ye
Shao-Lun Lee Valeriya Gribova Yifei Qi
Wei-Chiang Hong Guangchen Wang Yifei Sun
Hongyan Sang Hong Wang Yinglei Song
Jinhui Liu Haiyan Wang Ying Ling
Stephen Brierley Jingjing Wang Ying Shen
Haozhen Situ Ran Wang Yingying Qu
Sonja Sonja Waqas Haider Bangyal Lvjiang Yin
Jin-Xing Liu Pi-Jing Wei Yiping Liu
Haoxiang Zhang Fangping Wan Wenjie Yi
Sebastian Laskawiec Jue Wang Jianwei Yang
Shailendra Kumar Minghua Wan Yu-Jun Zheng
Junliang Shang Qiaoyan Wen Yonggang Lu
Guo Wei-Feng Takashi Kuremoto Yan Li
Yu-Bo Sheng Chuge Wu Yuannong Ye
Hongbo Shi Jibing Wu Yong Chen
Nobutaka Shimada Jinglong Wu Yongquan Zhou
Syeda Shira Moin Wei Wu Yong Zhang
Xingjia Lu Xiuli Wu Yuan Lin
Shoaib Malik Yahui Wu Yuansheng Liu
Feng Shu Wenyin Gong Bin Yu
Siqi Qiu Zhanjun Wang Fang Yu
Boyu Zhou Xiaobing Tang Kumar Yugandhar
Stefan Weigert Xiangfu Zou Liang Yu
Sameena Naaz Xuefeng Cui Yumin Nie
Sobia Pervaiz Lin Xia Xu Yu
Somnath Dey Taihong Xiao Yuyan Han
Sotanto Sotanto Xing Chen Yikuan Yu
Chao Wu Lining Xing Ying Wu
Yang Lei Jian Xiong Ying Xu
Surya Prakash Yi Xiong Zhiyong Wang
Wei Su Xiaoke Ma Shaofei Zang
Qi Li Guoliang Xu Chengxin Zhang
Hotaka Takizawa Bingxiang Xu Zehui Cao
FuZhou Tang Jianhua Xu Tao Zeng
Xiwei Tang Xin Xu Shuaifang Zhang
LiNa Chen Xuan Xiao Liye Zhang
Yao Tuozhong Takayoshi Yamashita Zhang Qinhu
Qing Tian Atsushi Yamashita Sai Zhang
Tianyi Zhou Yang Yang Sen Zhang
Junbin Fang Zhengyu Yang Shan Zhang
Wei Xie Ronggen Yang Shao Ling Zhang
Shikui Tu Yaolai Wang Wen Zhang
Umarani Jayaraman Yaping Yang Wei Zhao
Vahid Karimipour Yue Chen Bao Zhao
 ICIC 2018 Organization XIII

Zheng Tian Lida Zhu Quan Zou

Zheng Sijia Ping Zhu Qian Zhu
Zhenyu Xuan Qi Zhu Zunyan Xiong
Fangqing Zhao Zhong-Yuan Zhang Zeya Wang
Zhipeng Cai Ziding Zhang Yatong Zhou
Xing Zhou Junfei Zhao Shuyi Zhang
Xiong-Hui Zhou Juan Zou Zhongyi Zhou
 Contents – Part II

Deep Convolutional Neural Network for Fog Detection . . . . . . . . . . . . . . . . 1

Jun Zhang, Hui Lu, Yi Xia, Ting-Ting Han, Kai-Chao Miao,
Ye-Qing Yao, Cheng-Xiao Liu, Jian-Ping Zhou, Peng Chen,
and Bing Wang

Prediction of Crop Pests and Diseases in Cotton by Long Short Term

Memory Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Qingxin Xiao, Weilu Li, Peng Chen, and Bing Wang

Fault Diagnosis and Control of a KYB MMP4 Electro-Hydraulic

Actuator for LDVT Sensor Fault. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Tan Van Nguyen and Cheolkeun Ha

A Mask R-CNN Model with Improved Region Proposal Network

for Medical Ultrasound Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Jun Liu and PengFei Li

An Improved Double Hidden-Layer Variable Length Incremental Extreme

Learning Machine Based on Particle Swarm Optimization . . . . . . . . . . . . . . 34
Qiuwei Li, Fei Han, and Qinghua Ling

Breast Cancer Medical Image Analysis Based on Transfer

Learning Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Yi Liu and Xiaolong Zhang

A Fast Algorithm for Image Segmentation Based on Local

Chan Vese Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Le Zou, Liang-Tu Song, Xiao-Feng Wang, Yan-Ping Chen, Qiong Zhou,
Chen Zhang, and Xue-Fei Li

Multi-modal Plant Leaf Recognition Based on Centroid-Contour Distance

and Local Discriminant Canonical Correlation Analysis . . . . . . . . . . . . . . . . 61
Shanwen Zhang, Zhen Wang, and Yun Shi

Fine-Grained Recognition of Vegetable Images Based on Multi-scale

Convolution Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Xiu-Hong Yang, Ji-Xiang Du, Hong-Bo Zhang, and Wen-Tao Fan

Feature Analysis and Risk Assessment of Android Group

Based on Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Zhijie Xiao, Tao Li, and Yuqiao Wang
XVI Contents – Part II

Find the ‘Lost’ Cursor: A Comparative Experiment of Visually

Enhanced Cursor Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Chuanyi Liu and Rui Zhao

An Enhanced HAL-Based Pseudo Relevance Feedback Model in Clinical

Decision Support Retrieval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Min Pan, Yue Zhang, Tingting He, and Xingpeng Jiang

An RGBD Tracker Based on KCF Adaptively Handling

Long-Term Occlusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Xue-Fei Zhang, Ai-Ping Zeng, Shan Huang, Ming Qing, and Yi Zhou

Background Subtraction with Superpixel and k-means . . . . . . . . . . . . . . . . . 108

Yu-Qiu Chen, Zhan-Li Sun, Nan Wang, and Xin-Yuan Bao

Application of Embedded Database SQLite in Engine Fault

Diagnosis System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Li Huang and Hui Huang

Research on Stock Forecasting Based on GPU and Complex-Valued

Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Lina Jia, Bin Yang, and Wei Zhang

Development of Remote Monitoring System of Communication Base

Station Using IoT and Particle Filter Technology . . . . . . . . . . . . . . . . . . . . 129
Yang-Weon Lee

Prediction of Dissolved Oxygen Concentration in Sewage Using Support

Vector Regression Based on Fuzzy C-means Clustering . . . . . . . . . . . . . . . . 136
Xing-Liang Shi, Jian Zhou, Xiao-Feng Wang, and Le Zou

A Deep Reinforcement Learning Method for Self-driving. . . . . . . . . . . . . . . 143

Yong Fang and Jianfeng Gu

Image Classification Based on Improved Spatial Pyramid

Matching Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Li Feng, Xiaofeng Wang, and Dongfang Chen

A Novel ECOC Algorithm with Centroid Distance Based Soft

Coding Scheme. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Kaijie Feng, Kunhong Liu, and Beizhan Wang

A Compound Algorithm for Parameter Estimation of Frequency Hopping

Signal Based on STFT and Morlet Wavelet Transform . . . . . . . . . . . . . . . . 174
Bao-Lin Zhang, Jun Lv, and Jia-Rui Li
 Contents – Part II XVII

Recognition of Comparative Sentences from Online Reviews

Based on Multi-feature Item Combinations . . . . . . . . . . . . . . . . . . . . . . . . . 182
Jie Zhang, Liping Zheng, Lijuan Zheng, and Junyan Ge

Hierarchical Hybrid Code Networks for Task-Oriented Dialogue . . . . . . . . . . 194

Weiri Liang and Meng Yang

Characterization of Radiotherapy Sensitivity Genes by Comparative Gene

Set Enrichment Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Min Zhu, Xiaolai Li, Shujie Wang, Wei Guo, and Xueling Li

MiRNN: An Improved Prediction Model of MicroRNA Precursors

Using Gated Recurrent Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Meng Cao, Dancheng Li, Zhitao Lin, Cheng Niu, and Chen Ding

Using Deep Neural Network to Predict Drug Sensitivity of Cancer

Cell Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Yake Wang, Min Li, Ruiqing Zheng, Xinghua Shi, Yaohang Li,
Fangxiang Wu, and Jianxin Wang

PCM: A Pairwise Correlation Mining Package for Biological

Network Inference. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Hao Liang, Feiyang Gu, Chaohua Sheng, Qiong Duan, Bo Tian,
Jun Wu, Bo Xu, and Zengyou He

RP-FIRF: Prediction of Self-interacting Proteins Using Random Projection

Classifier Combining with Finite Impulse Response Filter . . . . . . . . . . . . . . 232
Zhan-Heng Chen, Zhu-Hong You, Li-Ping Li, Yan-Bin Wang,
and Xiao Li

NetCoffee2: A Novel Global Alignment Algorithm for Multiple PPI

Networks Based on Graph Feature Vectors. . . . . . . . . . . . . . . . . . . . . . . . . 241
Jialu Hu, Junhao He, Yiqun Gao, Yan Zheng, and Xuequn Shang

Dynamics of a Stochastic Virus Infection Model with Delayed

Immune Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Deshun Sun, Siyuan Chen, Fei Liu, and Jizhuang Fan

An Approach for Glaucoma Detection Based on the Features

Representation in Radon Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Beiji Zou, Qilin Chen, Rongchang Zhao, Pingbo Ouyang,
Chengzhang Zhu, and Xuanchu Duan

Screening of Pathological Gene in Breast Cancer Based

on Logistic Regression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
Yun Zhao and Xu-Qing Tang
XVIII Contents – Part II

Analysis of Disease Comorbidity Patterns in a Large-Scale

China Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Mengfei Guo, Yanan Yu, Tiancai Wen, Xiaoping Zhang, Baoyan Liu,
Jin Zhang, Runshun Zhang, Yanning Zhang, and Xuezhong Zhou

Stacking Multiple Molecular Fingerprints for Improving Ligand-Based

Virtual Screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Yusuke Matsuyama and Takashi Ishida

Combining mRNA, microRNA, Protein Expression Data and Driver Genes

Information for Identifying Cancer-Related MicroRNAs. . . . . . . . . . . . . . . . 289
Jiawei Lei, Shu-Lin Wang, and Jianwen Fang

ENSEMBLE-CNN: Predicting DNA Binding Sites in Protein Sequences

by an Ensemble Deep Learning Method. . . . . . . . . . . . . . . . . . . . . . . . . . . 301
Yongqing Zhang, Shaojie Qiao, Shengjie Ji, and Jiliu Zhou

Exploration and Exploitation of High Dimensional Biological Datasets

Using a Wrapper Approach Based on Strawberry Plant Algorithm . . . . . . . . 307
Edmundo Bonilla-Huerta, Roberto Morales-Caporal,
and M. Antonio Arjona-López

Prediction of LncRNA by Using Muitiple Feature Information Fusion

and Feature Selection Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
Jun Meng, Dingling Jiang, Zheng Chang, and Yushi Luan

Tri-Clustering Analysis for Dissecting Epigenetic Patterns Across

Multiple Cancer Types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Yanglan Gan, Zhiyuan Dong, Xia Zhang, and Guobing Zou

Efficient Framework for Predicting ncRNA-Protein Interactions

Based on Sequence Information by Deep Learning . . . . . . . . . . . . . . . . . . . 337
Zhao-Hui Zhan, Zhu-Hong You, Yong Zhou, Li-Ping Li,
and Zheng-Wei Li

BIOESOnet: A Tool for the Generation of Personalized Human Metabolic

Pathways from 23andMe Exome Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
Marzio Pennisi, Gabriele Forzano, Giulia Russo, Barbara Tomasello,
Marco Favetta, Marcella Renis, and Francesco Pappalardo

Accurate Prediction of Hot Spots with Greedy Gradient Boosting

Decision Tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
Haomin Gan, Jing Hu, Xiaolong Zhang, Qianqian Huang,
and Jiafu Zhao

A Novel Efficient Simulated Annealing Algorithm for the RNA Secondary

Structure Predicting with Pseudoknots . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Zhang Kai and Lv Yulin
 Contents – Part II XIX

TNSim: A Tumor Sequencing Data Simulator for Incorporating

Clonality Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
Yu Geng, Zhongmeng Zhao, Mingzhe Xu, Xuanping Zhang, Xiao Xiao,
and Jiayin Wang

Optimizing HP Model Using Reinforcement Learning . . . . . . . . . . . . . . . . . 383

Ru Yang, Hongjie Wu, Qiming Fu, Tao Ding, and Cheng Chen

Predicting Microbe-Disease Association by Kernelized Bayesian

Matrix Factorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
Sisi Chen, Dan Liu, Jia Zheng, Pingtao Chen, Xiaohua Hu,
and Xingpeng Jiang

Similarity-Based Integrated Method for Predicting

Drug-Disease Interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
Yan-Zhe Di, Peng Chen, and Chun-Hou Zheng

Nucleotide-Based Significance of Somatic Synonymous Mutations

for Pan-Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
Yannan Bin, Xiaojuan Wang, Qizhi Zhu, Pengbo Wen, and Junfeng Xia

Performance Analysis of Non-negative Matrix Factorization Methods

on TCGA Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
Mi-Xiao Hou, Jin-Xing Liu, Junliang Shang, Ying-Lian Gao,
Xiang-Zhen Kong, and Ling-Yun Dai

Identifying Characteristic Genes and Clustering via an Lp-Norm Robust

Feature Selection Method for Integrated Data . . . . . . . . . . . . . . . . . . . . . . . 419
Sha-Sha Wu, Mi-Xiao Hou, Jin-Xing Liu, Juan Wang, and Sha-Sha Yuan

Using Novel Convolutional Neural Networks Architecture to Predict

Drug-Target Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
ShanShan Hu, DeNan Xia, Peng Chen, and Bing Wang

Computational Prediction of Driver Missense Mutations in Melanoma . . . . . . 438

Haiyang Sun, Zhenyu Yue, Le Zhao, Junfeng Xia, Yannan Bin,
and Di Zhang

Further Evidence for Role of Promoter Polymorphisms in TNF Gene

in Alzheimer’s Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
Yannan Bin, Ling Shu, Qizhi Zhu, Huanhuan Zhu, and Junfeng Xia

Verifying TCM Syndrome Hypothesis Based on Improved Latent

Tree Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
Nian Zhou, Lingshan Zhou, Lili Peng, Bing Wang, Peng Chen,
and Jun Zhang
XX Contents – Part II

acsFSDPC: A Density-Based Automatic Clustering Algorithm

with an Adaptive Cuckoo Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 470
Chang Liu, Junliang Shang, Xuhui Zhu, Yan Sun, Jin-Xing Liu,
Chun-Hou Zheng, and Junying Zhang

Epileptic Seizure Detection Based on Time Domain Features and Weighted

Complex Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Hanyong Zhang, Qingfang Meng, Bo Meng, Mingmin Liu, and Yang Li

Multi-path 3D Convolution Neural Network for Automated Geographic

Atrophy Segmentation in SD-OCT Images . . . . . . . . . . . . . . . . . . . . . . . . . 493
Rongbin Xu, Sijie Niu, Kun Gao, and Yuehui Chen

Multi-information Fusion Based Mobile Attendance Scheme

with Face Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
Likai Dong, Qinlin Li, Tao Xu, Xuesong Sun, Dong Wang,
and Qingqing Yin

The Wide and Deep Flexible Neural Tree and Its Ensemble in Predicting
Long Non-coding RNA Subcellular Localization. . . . . . . . . . . . . . . . . . . . . 515
Jing Xu, Peng Wu, Yuehui Chen, Hussain Dawood, and Dong Wang

Research on Auto-Generating Test-Paper Model Based

on Spatial-Temporal Clustering Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 526
Yuling Fan, Likai Dong, Xuesong Sun, Dong Wang, Wang Qin,
and Cao Aizeng

A Semantic Context Model for Automatic Image Annotation . . . . . . . . . . . . 536

Xin Fu, Dong Wang, Sijie Niu, and Hengcai Zhang

Identifying Transportation Modes Using Gradient Boosting Decision Tree . . . 543

Xin Fu, Dong Wang, and Hengcai Zhang

A Multi-object Optimization Cloud Workflow Scheduling Algorithm Based

on Reinforcement Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 550
Wu Jiahao, Peng Zhiping, Cui Delong, Li Qirui, and He Jieguang

A AprioriAll Sequence Mining Algorithm Based on Learner Behavior. . . . . . 560

Zhenghong Yu and Dan Li

Emotion Recognition Based on Electroencephalogram Using a Multiple

Instance Learning Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570
Xiaowei Zhang, Yue Wang, Shengjie Zhao, Jinyong Liu, Jing Pan,
Jian Shen, and Tingzhen Ding

Using Hybrid Similarity-Based Collaborative Filtering Method

for Compound Activity Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 579
Jun Ma, Ruisheng Zhang, Yongna Yuan, and Zhili Zhao
 Contents – Part II XXI

Improving Initial Model Construction in Single Particle Cryo-EM

by Filtering Out Low Quality Projection Images . . . . . . . . . . . . . . . . . . . . . 589
Zhijuan Wang and Yonggang Lu

I-MMST: A New Task Scheduling Algorithm in Cloud Computing. . . . . . . . 601

Belal Ali Al-Maytami, Pingzhi Fan, and Abir Hussain

Recognition and Severity Rating of Parkinson’s Disease from Postural

and Kinematic Features During Gait Analysis with Microsoft Kinect . . . . . . . 613
Ilaria Bortone, Marco Giuseppe Quercia, Nicola Ieva,
Giacomo Donato Cascarano, Gianpaolo Francesco Trotta,
Sabina Ilaria Tatò, and Vitoantonio Bevilacqua

Rhino-Cyt: A System for Supporting the Rhinologist in the Analysis

of Nasal Cytology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 619
Giovanni Dimauro, Francesco Girardi, Matteo Gelardi,
Vitoantonio Bevilacqua, and Danilo Caivano

Recovering Segmentation Errors in Handwriting Recognition Systems . . . . . . 631

Claudio De Stefano, Francesco Fontanella, Angelo Marcelli,
Antonio Parziale, and Alessandra Scotto di Freca

A Deep Learning Approach for the Automatic Detection and Segmentation

in Autosomal Dominant Polycystic Kidney Disease Based on Magnetic
Resonance Images. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
Vitoantonio Bevilacqua, Antonio Brunetti, Giacomo Donato Cascarano,
Flavio Palmieri, Andrea Guerriero, and Marco Moschetta

A Model-Free Computer-Assisted Handwriting Analysis Exploiting

Optimal Topology ANNs on Biometric Signals in Parkinson’s
Disease Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 650
Vitoantonio Bevilacqua, Claudio Loconsole, Antonio Brunetti,
Giacomo Donato Cascarano, Antonio Lattarulo, Giacomo Losavio,
and Eugenio Di Sciascio

Particle Swarm Optimization with Convergence Speed Controller

for Sampling-Based Image Matting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 656
Yihui Liang, Han Huang, Zhaoquan Cai, and Liang Lv

iCorr-GAA Algorithm for Solving Complex Optimization Problem . . . . . . . . 669

Fangyuan Ding, Min Huang, Yongsheng Deng, and Han Huang

Two Possible Paradoxes in Numerical Comparisons

of Optimization Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681
Qunfeng Liu, Wei Chen, Yingying Cao, Yun Li, and Ling Wang
XXII Contents – Part II

Mosquito Host-Seeking Algorithm Based on Random Walk

and Game of Life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693
Yunxin Zhu, Xiang Feng, and Huiqun Yu

A Novel Energetic Ant Optimization Algorithm for Routing

Network Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 705
Xiang Feng and Hanyu Xu

A Comparison Study of Surrogate Model Based Preselection

in Evolutionary Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717
Hao Hao, Jinyuan Zhang, and Aimin Zhou

A Grouping Genetic Algorithm Based on the GES Local Search for Pickup
and Delivery Problem with Time Windows and LIFO Loading . . . . . . . . . . . 729
Feng Zhang, Bin Li, and Kun Qian

Evolutionary Structure Optimization of Convolutional Neural Networks

for Deployment on Resource Limited Systems . . . . . . . . . . . . . . . . . . . . . . 742
Qianyu Zhang, Bin Li, and Yi Wu

Research on Vehicle Routing Problem and Its Optimization Algorithm

Based on Assembled Building . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 754
Kun Jiang, Jun-qing Li, Ben Niu, Yongqin Jiang, Xiaoping Lin,
and Pei-yong Duan

Research on Vehicle Routing Problem with Time Windows Restrictions . . . . 763

Yun-Qi Han, Jun-Qing Li, Yong-Qin Jiang, Xing-Rui Chen, Kun Jiang,
Xiao-Ping Lin, and Pei-Yong Duan

Research on Swarm Intelligence Algorithm Based on Prefabricated

Construction Vehicle Routing Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 771
Xingrui Chen, Jun-Qing Li, Yongqin Jiang, Yunqi Han, Kun Jiang,
Xiaoping Lin, and Pei-Yong Duan

Improved Digital Password Authentication Method for Android System. . . . . 780

Bo Geng, Lina Ge, Qiuyue Wang, and Lijuan Wang

Steepest Descent Bat Algorithm for Solving Systems

of Non-linear Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 792
Gengyu Ge, Jiaxian Song, Juan Wang, and Aijia Ouyang

User Engagement Prediction Using Tweets. . . . . . . . . . . . . . . . . . . . . . . . . 802

Ameesha Mittal, Geetika Arora, Kamlesh Tiwari,
Vandana Dixit Kaushik, and Phalguni Gupta

Discrimination and Prediction of Protein-Protein Binding Affinity Using

Deep Learning Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809
Rahul Nikam, K. Yugandhar, and M. Michael Gromiha
 Contents – Part II XXIII

Principal Component Analysis-Based Unsupervised Feature Extraction

Applied to Single-Cell Gene Expression Analysis . . . . . . . . . . . . . . . . . . . . 816
Y-h. Taguchi

Finding Protein-Binding Nucleic Acid Sequences Using a Long Short-Term

Memory Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827
Jinho Im, Byungkyu Park, and Kyungsook Han

Constructing Gene Co-expression Networks for Prognosis

of Lung Adenocarcinoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
Byungkyu Park, Jinho Im, and Kyungsook Han

RMCL-ESA: A Novel Method to Detect Co-regulatory Functional

Modules in Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 840
Jiawei Luo and Ying Yin

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847

 Deep Convolutional Neural Network
for Fog Detection

Jun Zhang1(&), Hui Lu1, Yi Xia1, Ting-Ting Han1, Kai-Chao Miao2,

Ye-Qing Yao2, Cheng-Xiao Liu2, Jian-Ping Zhou2, Peng Chen3,
and Bing Wang4
1
School of Electronic Engineering and Automation,
Anhui University, Hefei 230601, Anhui, China
wwwzhangjun@163.com
2
Anhui Meteorological Bureau, Hefei 230031, Anhui, China
3
Institute of Health Sciences, Anhui University, Hefei 230601, Anhui, China
4
School of Electrical and Information Engineering,
Anhui University of Technology, Ma Anshan 243032, China

Abstract. Fog detection has becomes more and more important in recent years,
real-time monitoring information is very beneficial for people to arrange pro-
duction and life. In this paper, based on meterological satellite data (Himawari-8
standard data, HSD8), Covolutional Neural Network (CNN) is used to detect
fog. Since HSD8 consists of 16 channels, the original CNN is extended to
multiple channels for HSD8. Multiple Channels CNN (MCCNN) can make the
full exploitation of spatial and spectral information effectively. A dataset is
created from Anhui Area which consists of ground station data and grid data.
Different image sizes and convolutional kernels are used to validate the pro-
posed methods. The experimental results show that the proposed method
achieves 91.87% accuracy.

Keywords: Fog
CNN
Detection
Himawari-8

1 Introduction

Fog has becomes one of the major catastrophic weather. The formation of fog is caused
by a large number of tiny water droplets or ice crystals floating in the air near the
ground. Therefore, if the horizontal visibility is less than 1 km, it can cause disastrous
weather. In recent years, with the rapid development of socio-economic construction,
especially land, sea and air transportation, fog detection the has become more and more
important. Fog generally has the characteristics of rapid generation and development
[1]. Traditional ground-based observation are limited to the local areas, it can not meet
the requirements of large-scale and fast monitoring.
The earliest use methods of satellite remote sensing technology in fog monitoring is
visible light image cloud identification and dissipation. In the 1960s, with the devel-
opment of meteorological satellites, European and American countries started to use
the satellite remote sensing technology for fog identification [2]. Usually, the methods
of monitoring heavy fog mainly include two-channel method, principal component

© Springer International Publishing AG, part of Springer Nature 2018

D.-S. Huang et al. (Eds.): ICIC 2018, LNCS 10955, pp. 1–10, 2018.
https://doi.org/10.1007/978-3-319-95933-7_1
2 J. Zhang et al.

analysis method and texture processing method [3]. But now with the further research
on satellite remote sensing, more scholars use different spectral characteristics to study
the method of fog identification.
Although the previous method performs well, it still has some drawbacks. Most of
these methods are based on manual feature learning, so the performance of these
methods heavily depends on domain knowledge. Even satellite remote sensing data has
rich spectral information, which can not be fully used by these traditional approaches
[4]. However, deep learning is a effective method which can extract the features from
mass of data automatically. In the last few years, deep learning has been a powerful
machine learning technique for learning data-dependent and hierarchical feature rep-
resentations from raw data. Convolutional Neural Network (CNN) is an algorithm in
deep learning. CNN has been widely applied to image processing and computer vision
problems, such as image classification,image segmentation [5], action recognition and
object detection [6].
Himawari-8 Standard Data (HSD8) data has a very high spatial and spectral res-
olution. Every ten miniute, the satellite return 16 channel data. Such a large data set
make CNN framework more suitable and more effective in handling classification [7, 8]
problems. Deep learning has been widely used in hyperspectral image classification.
Hyperspectral image data usually have hundreds of spectral image. Similar to hyper-
spectral image data, HSD8 consist by 16 channel.
In this research, we propose CNN framework that classifies HSD8 data into mul-
tiple classes. The experimental result show that the proposed method achieve a better
classification performance.

2 First Section

2.1 CNN
Convolution Neural Network (CNN) is an efficient identification method that has been
developed and attracted a great deal of attention in recent years. Generally, the basic
structure of CNN includes two layers, one is a feature extraction layer, the input of each
neuron is connected with a local acceptance field of the previous layer. Once the local
feature is extracted, its location relationship with other features is also determined. The
second is the feature mapping layer. The feature mapping structure uses the sigmoid
function with small influence kernel as the activation function of the CNN, which
makes the feature map invariant to displacement. In addition, due to the shared weights
of neurons on a mapping surface, the number of free network parameters is reduced.
The special structure of CNN sharing local weight has unique advantages in speech
recognition [9] and image classification [10]. Weight sharing reduces the complexity of
the network. In particular, the features of multi-dimensional input vector images can be
directly input to the network avoid the complexity of data reconstruction in feature
extraction.
CNN is a feed forward neural network, the structure of the model is shown in the
Fig. 1. x is the input, h is the neuron output. When multiple cells are combined and
have a hierarchical structure, a neural network model is established. Neural network
 Deep Convolutional Neural Network for Fog Detection 3

connects many of these neurons into one network, the output of one neuron serves as
input to another neuron. Neural networks can have a wide variety of topologies. One of
the simplest is “multi-layer fully connected forward neural network.” Its input is
connected to every neuron on the first level of the network. The output of each neuron
in the previous layer is connected to the input of each neuron in the next layer. The
output of the last neuron is the output of the entire neural network.

x1
a1(2)

x2 a2(2)
a3(2)
x3 hw,b(x)

+1 +1

Layer L1 Layer L2 Layer L3

Fig. 1. Upper and lower neurons are all connected to the neural network

The order of CNN is the alternation of the convolutional layer with the pooling
layer, which mimics the complex and simple cell nature of the mammalian visual
cortex. Finally, the fully connected neural network ends. In ordinary deep neural
networks, neurons connect to all the neurons in the next level. CNN Unlike ordinary
neural networks, a typical CNN is as shown in Fig. 2. That is, each hidden activation is
calculated by multiplying the entire input x by the weight W in that layer. However, in
CNN, every hidden activation point h is obtained by inputting a small local weight W,
so that the weight W can be shared in the entire input space. Neurons belonging to the
same level share the same weight. The advantage of CNN weight sharing is that it helps
to reduce the total number of training parameters and leads to more efficient training
and more efficient models.
The pooling layer has two functions, the first effect is to retain the main features
while reducing the parameters (similar to the PCA [11]) and the amount of calculation
to prevent over-fitting and improve the model generalization ability. The second effect
is introduced of invariance. The pooling layer has the max pooling and the average
pooling, the typical pooling function is the max pooling. The max pooling divides the
input data into a set of non-overlapping windows and outputs the maximum for each
subregion, reducing the upper computational complexity and providing a form of
translational invariance. The computational chain of the CNN terminates in fully
connected network, the fully connected network contains the feature information
extracted by the convolutional and pooling layers, we mainly use it for classification. In
this article, we explore the appropriate architecture and strategy for CNN-based HSD8
classification.
4 J. Zhang et al.

Input x Convolutional layer Max pooling layer Fully connected layer

W
x1
h1
x2
h2
x3
h3
x4
h4
x5
x6 h5
h6
Shared weights

Fig. 2. A typical CNN architecture consisting of a convolutional layer, a max pooling layer, and
a fully connected layer.

2.2 CNN-Based HSD8 Classification

Architecture of the Proposed CNN Classifier
We know that CNN can be more competitive and perform better than humans in certain
visual issues and their ability to stimulate our research into the possibility of using
CNN for classification of HSD8 using spectral features. In this paper, we adopt the
CNN structure shown in Fig. 3. This network contains one input layer, two pairs of
convolution layers and max pooling layers, two fully connected layers and one output
layer in our network structure.

C1 M1
M C2 M2 F
F1 F2
Class Probability
Fully Connected Layer
Fully Connected Layer
Convolutional Later
Convolutional Later

Max Pooling Layer

Max Pooling Layer

Fig. 3. The network structure of the CNN used as our discriminative classifier

Training Strategies
Here, we introduce how to learn the parameter space of the proposed CNN classifier.
CNN’s training process is mainly divided into two steps: forward propagation and back
propagation [12]. The function of forward propagation is to calculate the actual clas-
sification result of the input data with the current parameters and the effect of back
propagation is to update the trainable parameters [13], so that the difference between
the actual classification output and the desired classification output as should be small
as possible.
 Deep Convolutional Neural Network for Fog Detection 5

Forward Propagation
As shown in the Fig. 3, the complete network consists of six layers, including the input
layer, the convolution layer C, the max pooling layer M, the full connection layer F.
We first define the problem related notations used throughout the paper. We assume
that h is all the training parameters, h ¼ fhig and i = 1, 2, 3, 4… where hi is the
parameter set between the ði  1Þ th and the i th layer. In this experiment we used a
6-layer convolution neural network. The output of the convolutional layer is as follows:

xi þ 1 ¼ fi ðui Þ ð1Þ

ui ¼ hTi xi þ bi ð2Þ

That xi is the input of the i th layer, h is weight. hTi is a weight matrix of the i th layer
acting on the input data, and bi is an additive bias vector for the i th layer. f ðÞ is the
activation function. The activation function is to introduce nonlinear factors. The
commonly used activation functions are Tanh, Sigmoid, Relu. In our designed archi-
tecture, we choose the hyperbolic tangent function Relu as the activation function. The
maximum function max(u) is used in layer M.
We want to use the hypothetical function to estimate the probability value pðy ¼ jjxÞ
for each category j. That is, we want to estimate the probability of each classification
result for x occurring. Therefore, our hypothetical function will output a k-dimensional
vector to represent the probabilities of these K estimates. Specifically, our hypothetical
function takes the following form, the softmax regression model is defined as:
2 3 2 T ðiÞ
3
PðyðiÞ ¼ 1jh; x; bÞ eh 1 x þ b1

6 7 6 T ðiÞ 7
6 PðyðiÞ ¼ 2jh; x; bÞ 7 6 eh 2 x þ b 2 7
ðiÞ 6 7 1 6 7
yj ¼6 ¼ Pk 6 7 ð3Þ
4 . . .: :. . . 75 j¼1 e hTj xðiÞ þ bj 6
4 . . . 7
5
PðyðiÞ ¼ 2jh; x; bÞ
T ðiÞ
ehj x þ bj

Back Propagation
In the back propagation stage, the backpropagation algorithm gives an efficient way to
use the gradient descent algorithm on all parameters, so that the loss function of the
neural network model on the training data is as small as possible. Backpropagation
algorithm is the core algorithm of training neural network, which can optimize the
parameters of neural network according to the defined loss function, so that the loss
function of neural network model on training data set reaches a smaller value. The
optimization of the parameters in the neural network model directly determines the
quality of the model, which is a very important step when using the neural network.
The loss function used in this work is defined as:

1X m X K n o
ðiÞ
JðhÞ ¼  1 j ¼ Y ðiÞ logðyj Þ ð4Þ
m i¼1 j¼1
6 J. Zhang et al.

ðiÞ
where m is the number of training samples. Y is the desired output. yj is the j th value
of the actual output yðiÞ of the i th training sample and is a vector whose size is K. In the
desired output Y ðiÞ of the i th sample, the probability value of the labeled class is 1, and
 
the probability values of other classes are 0.1 j ¼ Y ðiÞ means, if j is equal to the
desired label of the i th training sample, its value is 1; otherwise, its value is 0.
In the training process, we need to get the prediction label by finding the maximum
value in the output vector, which requires that we actually output is closer to the
expectation. When the difference between them is small enough, the iteration stops, at
which point the actual output is closer to the desired output. Experiments show that as
the number of training iterations increases, the return value of the loss function
becomes smaller, which can meet our needs. Finally, the trained CNN is ready for
HSD8 classification. Training samples are first randomly divided into batches, each
batch containing the same number of samples. The batch size has a certain impact on
the experimental results. During training, we trained CNN using a stochastic gradient
descent, each iteration only one batch was sent to the network for training. When the
maximum number of iterations is reached, the training process will stop, then the
training model will be output. In the testing process, we first input the test samples into
the training network [14]. The parameters of the network are obtained by using the
trained network parameters and then finding the maximum value in the output vector.
The training and testing process is shown in Fig. 4.

Batch 1
Training
Batch 2
Data & CNN
Label Training

Batch M

CNN Predicted
Test Test Data
Model Label

Fig. 4. Illustration of the parameter process

3 Experiments and Discussion

The Data Sets: This experiment using HSD8, total of 16 channels. When the visibility
is greater than 1000 m, it is negative sample (without fog) and the other sample is
positive sample (with fog).
Impact of parameter settings: The parameter selection is very important to the
performance of the network. Such as the depth of the network, the size and number of
convolution kernels, the batch size and the learning rate. In this experiment, we
evaluated in detail the effect of different parameter settings on the sensitivity of the
Another random document with
no related content on Scribd:
que solo el hidalgo se podia en
aquella tierra mantener, que el
labrador pechero era neçesario
morir de hanbre.
Miçilo.—¿Pues porque no se iba
tu padre á vibir a otra tierra?
Gallo.—Son tan acobardados
para en eso los labradores, que
nunca se atreuen a hazer
mudança de la tierra donde
naçen: porque vna legua de sus
lugares les pareçe que son las
Indias: y imaginan que ay alla
gentes que comen los honbres
biuos. Y por tanto muere cada
vno en el pajar donde naçio,
avnque sea de hanbre. Y deste
padre naçimos dos hijos varones,
de los quales yo fue el mayor,
llamado por nonbre Alexandro. Y
como vimos tanta miseria como
passauan con el señor los
labradores, pensauamos que si
tomauamos offiçios que por
entonçes nos libertassen se
oluidaria nuestra vileza, y
nuestros hijos serian tenidos y
estimados por hydalgos y viuirian
en libertad. Y ansi yo elegi ser
saçerdote, que es gente sin ley; y
mi hermano fue herrero, que en
aquella tierra son los herreros
exentos de los pedidos, pechos y
velas del lugar donde siruen la
ferreria. Y ansi yo demandé
liçencia a mi padre para aprender
a leer: y avn se le hizo de mal
porque le seruia de guardar vnos
patos. y ojear los pajaros que no
comiessen la simiente de vn linar.
En conclusion mi padre me
encomendo[362] por criado y
monaçino de vn capellan que
seruia vn beneffiçio tres leguas de
alli. ¡O Dios omnipotente, quien te
dixera las bajezas y poquedades
deste honbre! Por cierto si yo no
huuiera tomado la mano oy para
te contar[363] de mi y no de otros,
yo te dixera cosas de gran
donayre. Pero quierote hazer
saber que ninguno dellos sabe
más leer que deletrear y lo que
escriben aslo de sacar por
discreçion. En ninguna cosa estos
capellanes muestran ser
auentajados, sino en comer y
beber: en lo qual no guardan
tiempo ni medida ni razon. Con
este estuue dos años que no me
enseñó sino a mal hazer, y mal
dezir, y mal pensar y mal
perseuerar. A leer me enseñó lo
que el sabia, que era harto poco,
y á escreuir vna letra que no
pareçia sino que era arado el
papel con pies de escarabajos. Ya
yo era buen moço de quinze
años, y entendia que para yo no
ser tan asno como mi amo que
deuia de saber algun latin. Y ansi
me fue á Zamora a estudiar
alguna gramatica: donde llegado
me presenté ante el bachiller y le
dixe mi necesidad, y el me
preguntó si traya libro: y yo le
mostré vn arte de gramatica que
auia hurtado a mi amo, que fue de
los de Pastrana que auia mas de
mil años que se inprimió. Y el me
mostró en el los nominatiuos que
auia de estudiar.
Miçilo.—¿De qué te mantenias?
Gallo.—Dauame el bachiller los
domingos vna çedula suya para
vn cura, o capellan de vna aldea
comarcana el qual me daua el
çetre del agua bendita los
domingos y andaua por todas las
casas a la hora del comer
echando a todos agua: y en cada
casa me dauan vn pedaço de
pan, con los quales mendrugos
me mantenia en el estudio toda la
semana. Aqui estube dos años:
en los quales aprendi
declinaciones y conjugaçiones:
genero, preteritos y supinos. Y
porque semejantes honbres
que[364] yo luego nos
enhastiamos de saber cosas
buenas, y porque nuestra
intinçion no es saber más: sino
tener alguna noticia de las cosas
y mostrar que emos entendido en
ello quando al tomar de las
ordenes nos quisieren examinar.
Porque si nuestra intinçion fuesse
saber algo perseuerariamos en el
estudio. Pero en ordenandonos
començamos a oluidar y damonos
tan buena priesa que si llegamos
a las ordenes neçios, dentro de
vn mes somos confirmados
asnos. Y ansi me sali de Çamora,
donde estudiaua harto de mi
espaçio, y por estar ya enseñado
á mendigar con el çetre sabiame
como miel el pedir: y por tanto me
bolui a ello[365]. Y ansi acordé de
yrme por el mundo en compañia
de otros perdidos como yo, que
luego nos hallamos vnos a otros.
Y en esta compañia fue gran
tiempo zarlo, ó espinel: y alcançe
en esta arte de la zarleria todo lo
que se pudo alcançar.
Miçilo.—Nunca esa arte á mi
noticia llegó: declarate me mas.
Gallo.—Pues quiero
descubrirtelo todo de raiz. Tu
sabras que yo tenia la persona de
estatura creçida y andaua vestido
en diuersas prouinçias de
diuersos atauios, porque ninguno
pudiesse con mala intinçion
aferrar en mi. Pero mas á la
contina traya vna vestidura de
vuriel algo leonado obscuro,
honesta, larga y con vna barua
espesa y muy prolixa, de grande
autoridad y un manteo encima,
puesto á los pechos vn boton[366].
Otras vezes mudando las tierras
mudaua el vestido: y con la
mesma barua vsaua de vn habito
que en muchas prouinçias llaman
veguino: con vna saya y vn
escapulario de Religioso que
hazia vida en la soledad de la
montaña; vna cayada y vn rosario
largo, de vnas cuentas muy
gruesas en la mano, que cada
vez que la vna cuenta caya sobre
la otra lo oyan todos quantos en
vn gran templo estuuiessen.
Publiqué adiuinar lo que estaua
por venir, hallar los perdidos,
reconçiliar enamorados, descubrir
los ladrones, manifestar los
thesoros, dar remedio façil á los
enfermos y avn resuçitar los
muertos. Y como de mí los
honbres tenian noticia venian
luego prostrados con mucha
humildad a me adorar y bessar
los pies y a ofreçerme todas sus
haziendas, llamandome todos
propheta y diçipulo y sieruo de
Dios, y luego les ponia en las
manos vno versos que en vna
tabla yo traya scriptos con letras
de oro sobre vn barniz negro; que
dezian de esta manera:

Muneribus decorare meum

vatem atque ministrum
precipio: nec opum mihi cura,
at maxima vatis.
Estos versos dezia yo auermelos
enbiado Dios con vn angel del
çielo, para que por[367] su
mandado fuesse yo de todos
honrrado y agradeçido como
ministro y sieruo de su diuina
magestad. Hallé por el reyno de
Portogal y Castilla infinitos
honbres y mugeres los quales
avnque fuessen muy ricos y de
los más prinçipales de su
republica, pero eran tan tímidos
superstiçiosos que no alçauan los
ojos del suelo sin escrupulizar.
Eran tan façiles en el credito que
con vna piedra[368] arrebuxada en
unos trapos ó vn pergamino con
vnos plomos ó sellos colgando,
en las manos de vn hombre
desnudo y descalço luego se
arrojauan y humillauan al suelo, y
venian adorando y ofreciendose a
Dios sin se leuantar de alli hasta
que el prestigioso questor los
leuantasse con su propria mano;
y ansi estos como me vian con
aquella mi santidad vulpina
façilmente se me rendian sin
poder resistir. Venian á consultar
en sus cosas conmigo todo lo que
deuian, ó querian hazer y yo les
dezia, que lo consultaria con Dios,
y que yo les responderia su diuina
determinacion, y ansi a sus
preguntas procuraua yo
responder con gran miramiento
porque no fuesse tomado en
palabras por falso y perdiesse el
credito. Sienpre daua las
respuestas dubdosas, ó con
diuersos entendimientos, sin
nunca responder absolutamente a
su intinçion. Como a vno que me
preguntó; qué preçeptor daria a
vn hijo suyo que le queria poner al
estudio de las letras. Respondi
que le diesse por preçeptores al
Antonio de Nebrija y a Sancto
Thomas. Dando á entender que le
hiziesse estudiar aquellos dos
auctores, el vno en la gramatica y
el otro en la theología; y suçedió
morirse el mochacho dentro de
ocho dias; y como sus amigos
burlasen del padre porque daua
credito a mis desuarios y de mis
juizios llamandolos falsos,
respondió que muy bien me auia
yo dicho: porque sabiendo yo que
se auia de morir, di a entender
que auia de tener por preçeptores
aquellos allá. Y a otro que auia de
hacer vn camino y temiasse de
vnos enemigos que tenia, que me
preguntó si le estaua bien yr
aquel camino. Respondi que más
seguro se estaua en su casa si le
podia escusar; y caminó por
burlar[369] de mi juizio, y sucedió
que salieron sus enemigos y
hirieronle mal. Despues como
aquel juizio se publicó me valio
muchos dineros a mi: porque
desde allí adelante no auian de
hazer cosa que no la viniessen
comigo á consultar pagandomelo
bien. En fin en esta manera dy
muchos y diuersos juizios que te
quisiera agora contar, sino fuera
porque me queda mucho por
dezir. Deziamos yo ser Juan de
vota Dios[370].
Miçilo.—¿Qué hombre es ese?
Gallo.—Este fingen los zarlos
superstiçiosos vagabundos que
era vn zapatero que estaua en la
calle de amargura en Hierusalen,
y que al tiempo que passauan a
Cristo presso por aquella calle,
salió dando golpes con vna horma
sobre el tablero diziendo: vaya,
vaya el hijo de María; y que Cristo
le auia respondido: yo yré y tú
quedarás para sienpre jamas para
dar testimonio de mi; y para en fe
desto mostraua yo vna horma
señalada en el braço, que yo
hazia con cierto artifiçio muy
façilmente, que pareçia estar
naturalmente empremida allí: y a
la contina traya vn compañero del
mesmo offiçio y perdiçion que
fuesse mas viejo que yo, porque
descubriendonos el vno al otro lo
que en secreto y confession con
las gentes tratauamos,
pareçiendo vn dia el vno y otro dia
el otro les mostrauamos tener
speçie de divinaçion y spiritu de
profeçia, lo qual sienpre nosotros
queriamos dar á entender. Y
haziamos se lo façilmente creer
por variarnos cada dia en la
representaçion; y deziales yo que
en viendome viejo me yba a
bañar al rio Xordan y luego boluia
de edad de treynta y tres años
que era la edad en que Cristo
murio. Otras vezes dezia que era
vn peregrino de Hierusalen,
honbre de Dios, enviado por él
para declarar y absoluer los
muchos pecados que auia[371]
secretos en el mundo, que por
verguença los honbres no los
osan descubrir ni confesar a
ningun confessor.
Miçilo.—¿Pues para qué era
eso?
Gallo.—Porque luego en
auiendoles hecho creer que yo
era qualquiera destos dos
façilmente los podia abunir a
qualquiera cosa que los quisiesse
sacar. Luego como los tenia en
este estado començaua la zarleria
cantandoles el espinela, que es
vn genero de diuinança, a manera
de dezir la buenauentura. Es vna
agudeça y desenboltura de
hablar, con la qual los que
estamos platicos en ello sacamos
façilmente qualesquier genero de
scollos (que son los pecados) que
nunca por abominables se
confessaron a saçerdote. En
començando yo a escantar con
esta arte luego ellos se
descubren.
Miçilo.—Yo querria saber qué
genero de pecados son los que
se descubren a ti por esta arte, y
no al saçerdote?
Gallo.—Hallaua mugeres que
tuuieron açeso con sus padres,
hijos y con muy çercanos
parientes, y vnas mugeres con
otras con instrumentos hechos
para effectuar este viçio; y otras
maneras que es verguença de las
dezir; y hallaua honbres que se
me confessauan auer cometido
grandes inçestos, y con animales
brutos, que por no infiçionar el
ayre no te los quiero contar. Son
estos pecados tan abominables
que de pura verguença y miedo
honbres ni mugeres no los osan
fiar ny descubrir a sus curas ni
confessores; y ansi aconteçe
muchos[372] destos neçios
morirse sin nunca los confessar.
Miçilo.—Pues de presumir es
que muchos destos honbres y
mugeres, pensando bastar
confessarlos a ti se quedaron sin
nunca á sacerdote los confessar.
Gallo.—Pues ese es vn daño
que trae consigo esta peruersa
manera de vibir, el qual no es
daño qualquiera sino de gran
caudal.
Miçilo.—Querria saber de ti, qué
virtud, o fuerça tiene esa arte que
se los hazeis vosotros confessar,
y qué palabras les dezis?
Gallo.—Fuerça de virtud no es:
pero antes industria de Sathanas.
La manera de palabras era: que
luego les dezia yo que por auer
aquella persona naçido en vn dia
de vna gran fiesta en çinco puntos
de Mercurio y otros çinco de
Mars, por esta causa su ventura
estaua en dos puntos de gran
peligro, y que el vn punto era
vibo, y el otro era muerto, y que
este punto vibo conuenia que se
cortasse, porque era vn gran
pecado que nunca confessó, por
el qual corria gran peligro en la
vida. En tanta manera que si no
fuera porque Dios le quiso
guardar por los ruegos del
bienauenturado San Pedro, que
era mucho su abogado ante Dios,
que muchas vezes le ha cometido
el demonio en grandes, afrentas
donde le quiso auer traydo a la
muerte; y que agora era enbiado
por Dios este su peregrino de
Hierusalen y santo profeta; que
soy vno de los doze peregrinos
que residen á la contina en el
sancto sepulcro de Hierusalen en
lugar de los doze apostoles de
Cristo; y que yo soy su abogado
San Pedro que conuiene que el
me le aya de descubrir y
confessar para que yo se le
absuelua, y avn pagarle[373] por
el, y asegurarle que no penará ni
peligrará por el[374] pecado más.
Y ansi él luego me descubre su
pecado por graue y inorme que
sea; y prostrado por el suelo
llorando me pide misericordia y
remedio y le mande quanto yo
quisiere que haga para ser
absuelto, que en todo me
obedeçerá y avn me dará quanto
yo le pidiere y el tuuiere para su
neçesidad; y ansi quando yo veo
a la tal persona tan obediente y
rendida digola. Pues mira,
hermana, que este pecado se ha
de absoluer con tres signos y tres
cruzes y tres psalmos y tres
misas solenes: las quales se han
de dezir en el templo del Santo
Sepulcro de Hierusalen, y que
son misas de mucha costa y
trabajo, porque las han de dezir
tres cardenales y rebestirse con
ellos al altar tres obispos; y
hanlas de offiçiar tres patriarcas
vestidos de pontifical, y han de
arder alli tres çirios a cada misa,
que pesse cada vno seys libras
de cera; y luego dize el tal
penitente:
Pues vos mi padre y santo señor
vays allá hazedlas dezir, y yo al
presente daré los dineros y
limosna que pudiere y boluiendo
vos por aqui lo acabaré de pagar;
y yo respondo: que a mi me
conuiene forçado estar en
Hierusalen la Semana Santa, y
que en llegando se las haré dezir,
y ansi luego el penitente me da
diez y veinte[375] ducados y más,
o menos como cada qual tiene la
facultad, y yo la doy vna señal por
la qual quedo de boluer a la visitar
dentro de vn año o dos, sin
pensarla mas ver; y otras vezes
para auctoriçar esta mi mala arte
digoles: que yo le daré parte del
gran trabajo que tengo de reçebir
en el camino que emos de hazer
los escolares peregrinos de
Hierusalen quando todos juntos
vamos la Santa pasqua de
Resureçion por el olio y crisma a
la torre de Babilonia, como lo
tenemos por costunbre y promesa
traerlo nosotros doze para la
iglesia de Dios; lo qual se trae en
doze cauallos yendo nosotros a
pie. Que van luego los siete y
quedan los çinco aguardando; y
aquellos siete que van lleuan
siete ropas ricas y siete armas,
con las quales peleamos con
siete gigantes que guardan el
santo crisma y el olio de noche y
de dia, y como son mas fuertes
que nosotros dannos grandes
palos y bofetadas, hasta que
vienen del çielo siete donzellas en
siete nubes y en su fabor siete
estrellas; las quales peleando con
los gigantes los vencen y ansi las
damos las siete ropas, y nos
cargan los cauallos del Santo olio
y crisma y nos venimos con ello á
Hierusalen para que en la Santa
pascua de Resurreçion se
distribuya por toda la cristiandad;
y ansi por la misericordia de Dios
nuestro señor, por esta tu limosna
te haré parçionera deste trabajo
que en este viaje tengo de lleuar
por la iglesia de Dios; y demas
desto porque quedes más
purgada deste pecado me vañaré
por ti en la fuente y rio Xordan
vna vez. Y con este fingimiento y
enbaymiento, fiçiones y engaños
las hazia tan obedientes a mi
mandado, que despues de
auerme dado su hazienda si
queria tenia açesso con ellas a
medida de mi voluntad, y ellas se
preçiaban auer tenido açesso con
el profeta diçipulo de Dios y
peregrino[376] santo de
Hierusalen, sieruo de Jesu-
Cristo[377]. Y se tenian por muy
dichosos los maridos por auer
querido yo ansi bendezir a su
muger; y ellas se piensan quedar
benditas para sienpre jamas con
semejantes bendiçiones. En estas
maldades querria yo mucho que
el mundo estuviesse auisado, y
que no diesse lugar ninguno a se
dexar engañar de semejantes
honbres malos, pues todo esto es
manifiesta mentira y fiçion.
Y sé yo que al presente andan
muchos por el mundo, los quales
tienen engañada la mayor parte
de los cristianos, y se debria
procurar que los juezes los
vuscassen, y hallados los
castigassen en las vidas, porque
es vna speçie de superstiçion y
hurto el mas nefando que entre
infieles nunca se vsó, ni se sufrió.
Y porque veas quanta es la
desverguença y poquedad de los
semejantes hombres te quiero
contar vn passo que passé,
porque entiendas que los tales
ninguna vellaqueria ni poquedad
dexan de acometer y executar.
Sabras que vn dia yuamos tres
compañeros del offiçio del zarlo y
espinela, que andauamos
vuscando nuestra ventura por el
mundo. Y como llegamos acaso
en vna çiudad á la hora del
comer, nos entramos en vn
bodegon, donde comimos y
bebimos muy a pasto todos tres, y
acordamos que se saliesse el vno
á vuscar çierto menester, y como
se tardasse algo fuele el otro
vuscar: y ansi me dexaron solo a
mi por gran pieza de tiempo, y
dixome la bodegonera: hermano,
pagad, ¿que aguardais?
Respondi yo: aguardo aquellos
compañeros que fueron á vuscar
çierta cosa para nuestra
necesidad; y ella me dixo: pagad
que por demas los esperais: por
neçios los ternia si ellos
boluiessen acá; y yo le pregunté
quanta costa estaua hecha, para
pagarla; y ella contando á su
voluntad y sin contradiçion dixo
que quatro reales auiamos
comido y bebido; y luego me
leuanté de la mesa viniendome
para la puerta de la casa
mostrando vuscar la bolsa para la
pagar, y dixela: señora echadme
en vna copa vna vez de vino, que
todo junto lo pagaré: y diziendo
esto nos fuemos llegando a vn
cuero de vino que sobre vna
mesa tenia junto a la[378] puerta, y
la buena dueña, avnque no era
menos curial en semejantes
maldades que yo, descuydose: y
desató luego el cuero echando la
cuerda sobre el hombro por tener
con la vna mano el piezgo y con
la otra la medida, y começando
ella a medir le tomé yo la cuerda
del ombro y fueme lo mas
solapadamente que yo pude por
la calle adelante y avnque ella me
llamaua no le respondia: ni ella
por no dexar el cuero desatado
me vio mas hasta oy. Cansado ya
desta miserable y trabajada vida
fueme a ordenar para clerigo.
Miçilo.—¿Con que letras te yuas
al examen?
Gallo.—Con seys conejos y
otras tantas perdiçes que lleué al
prouisor, y ansi maxcando vn
euangelio que me dio a leer, y
declinando al reues vn nominatiuo
me passó, y al escrivano que le
dixo que no me deuia de ordenar
respondio: andad que es pobre y
no tiene de qué viuir.
Miçilo.—Por çierto que todo va
ansi. Que yo conozco clerigos tan
neçios y tan desuenturados que
no les fiaria la tauerna del lugar.
No saben sino coger la pitança y
andar, y si les preguntais, ¿donde
vays tan apriesa? Responde él
con el mesmo desasosiego: a
dezir misa. ¿Que no ay mas? Por
vn miserable estipendio, que si no
fuesse por él no la diria.
Gallo.—La cosa que más
lastimado me tiene el coraçon en
las cosas de la cristiandad es
esta: el poco acatamiento que
tienen estos capellanes en el
dezir misa. Que de todas las
naçiones del mundo no ay
ninguna que más bienes aya
reçebido de su Dios que los
cristianos: que los de los otros no
son dioses: no los pueden dar
nada; y con tantas merçedes
como los ha hecho, que avn asi
mesmo se les dio, y no ay naçion
en el mundo que menos
acatamiento tenga á su Dios que
los cristianos: y por eso les da
Dios enfermedades, pestelençias,
hambres, guerras, herejes. Que
en vn rincon de la cristiandad ay
todos estos males y justamente
los mereçen. Que como ellos
tratan a Dios ansi los trata él a
ellos a osadas. Que vno que para
tauernero no es sufiçiente se
haze saçerdote por ganar de
comer: y tanbien tienen desto
gran culpa los seglares, por el
trato que anda de misas y varatos
malos: que si esto no huuiesse no
se ordenaria tanto perdido y
oçioso como se ordenan con
confiança desto. Escriben los
historiadores por gran cosa, que
vn papa ordenó tres saçerdotes y
çinco diaconos, y ocho
subdiaconos. Y agora no hay
obispo de anillo que cada año no
aya ordenado quinientos desos
ydiotas y mal comedidos asnos.
Por eso determinó la iglesia que
los saçerdotes no se pudiessen
ordenar sino en qvatro temporas:
porque entonçes ayunasse el
pueblo aquellos dias, y rogassen
á Dios que les diesse buenos
saçerdotes, y por yr en ello tanta
parte del bien de la republica.
Pues y crees tú que se haze esto
alguna vez? Yo confio que nunca
le passa por pensamiento mirar
en esto a honbre de toda la
cristiandad: ni avn creo que nunca
tú oyste esto hasta agora.
Miçilo.—No por çierto.
Gallo.—Pues sabete que es la
verdad. Aveis de rogar a Dios que
os dé buenos saçerdotes: porque
algunos saçerdotes ay que no os
los dio Dios, sino el demonio, la
simonia y avariçia. Como a mí
que en la verdad yo me ordené
por auariçia de tener de comer: y
simoniacamente me dieron las
ordenes por seys conejos y otras
tantas[379] perdiçes, y permitelo
Dios, Quia qualis populus talis est
sacerdos. Quiere Dios daros
ruynes saçerdotes por los
pecados del pueblo: porque qual
es el pueblo tales son sus[380]
sacerdotes.
Miçilo.—Por çierto que en quanto
dizes has dicho verdad, y que me
he holgado mucho en oyrte.
Boluamos, pues, a donde
dexaste: porque quiero saber tú
que tal saçerdote heziste.
Gallo.—Por çierto dese mesmo
jaez: y avn peor que todos los
otros de que emos hablado.
Luego como fue saçerdote el
primer año mostré gran santidad:
y çertificote que yo mudé muy
poquito de mi vida passada: pero
mostraua gran religion: y ansi vibi
dos años aqui en esta villa: y
como me viessen la bondad que
yo representaua, que siempre
andaua en compañía de vna trulla
de clerigos santos que ha auido
de pocos tiempos en ella,
andando a la cortina visitando los
hospitales y corrales donde
auia[381] pobres, en compañía de
vnas mugerçillas andariegas y
vagarosas, callegeras que no
sufren estar vn momento en sus
casas quedas, que estas con todo
desassosiego tratauan en la
mesma santidad.
Miçilo.—Mayor santidad tuuieran
estando en sus casas en oraçion
y recogimiento.
Gallo.—-De las quales[382]
teniamos nuestras çiertas
granjerias, como camisas,
pañizuelos de narizes: y la ropa
blanca labada cada semana: y
algunas ollas y otros guisadillos
regalados[383] y algunoe
vizcochos y rosquillas: y como
vian todos la bondad que
representaua hablome vn letrado
rico si queria enseñarle vnos
niños pequeños que tenia, sus
hijos.