Synopsis

“ CROWD MONITORING SYSTEM USING DEEP LEARNING”

Abstract:
 The understanding of crowd behavior in public spaces is an important part of the design of
new pedestrian facilities, for major layout modifications to existing areas and for the daily
management of sites subject to crowd traffic.

 This article presents some image processing techniques which, using existing closed-circuit
television systems, can support both data collection and on-line monitoring of crowds. The
application of these methods could lead to a better understanding of crowd behavior,
improved design of the built environment and increased pedestrian safety.

Introduction:
 Human identification and monitoring are not an easy task in the case of a large and densely
populated crowd. Existing crowd monitoring systems work well for individuals and not for
mass gatherings as a whole.

 Maintaining public order in certain crowded places such as airports, carnivals, sports events,
and railway stations is very essential. In crowd management system, counting people is an
essential factor. Particularly in smaller areas, increase in the number of people create
problems such as fatalities, physical injury etc.

Literature survey:
Sr. No. Reference Work Description Problem found
Name

1 Estimation of Extracts crowd Output is given in terms of range

crowd density density features
using Image from digitized
processing images

2 A Systematic Localization Use of h/w can increase system cost

Review of schemes in IoT
Location
Aware
Schemes in the
Internet of
Things
 3 Image Generation and Background pixels for generating a
Processing calibration to background image without pedestrians.
Techniques for enhance the
Crowd Density previously-
Estimation developed method of
Using a crowd density
Reference estimation using a
Image. reference image.

4 CNN-based Development of Accurate estimation of number of object is

Density techniques for not possible.
Estimation and counting crowd.
Crowd
Counting: A
Survey

5 SegNet: A Deep Fully-convolutional Decoders are required for topological.

Convolution neutral network
Encoder- architecture for
Decoder semantic pixel-wise
Architecture for segmentation
Image termed SegNet.
Segmentation

System Architecture:-
Objective:
• To manage the details of Accounts,Internet Banking
•  To attract customer And helping the customers by providing online services
• To making the banking system easier.

Methodology:
• Convolutional Neural Network - A Convolutional Neural Network (CNN) is a Deep Learning
algorithm which can take in an input image, assign importance (learnable weights and biases)
to various aspects/objects in the image and be able to differentiate one from the other.
• CNNs use image recognition and classification in order to detect objects, recognize faces, etc.
They are made up of neurons with learnable weights and biases. CNNs are primarily used to
classify images, cluster them by similarities, and then perform object recognition.

• Machine Learning
 • Machine learning is a method of data analysis that automates analytical model building. It is a
branch of artificial intelligence based on the idea that systems can learn from data, identify
patterns and make decisions with minimal human intervention.

• Deep Learning
• Deep learning is part of a broader family of machine learning methods based on artificial
neural networks with representation learning. Learning can be supervised, semi-supervised or
unsupervised.

Conclusion:
 In this seminar we examined real time crowd monitoring, basic system architecture and ways
to monitor, capture and process the crowd images.
 This examined work would be implemented in our final year project as to monitor various
suspicious activities in public areas for better security and safety.

References:
[1] Tarun Kulshrestha, Divya Saxena, Member, IEEE, Rajdeep Niyogi, Jiannong Cao Fellow, IEEE,
“Real-time Crowd Monitoring using Seamless Indoor-Outdoor Localization,”  IEEE Transaction on
Mobile Computing Volume: 19, Issue: 3, March 1 2020.
[2] “Estimation of crowd density uses image processing technology”: - A.N. Maranal, S. A. Velastin2,
L. F. Costa3 and R. A. Lotufo4.
[3] “Crowd Monitoring Using Image Processing:” - by Anthony C. Davies, JiaHong Yin and Sergio
A. Velastin.
[4] Chouchang Yang and Huai-Rong Shao, “Wi-Fi-Based Indoor Positioning,” IEEE
Communications Magazine, March 2015.
[5] Muneeb A. Khan, Abdul Saboor, Hyun-chul Kim, Heemin Park, “A Systematic Review of
Location Aware Schemes in the Internet of Things,’’ ISSN: 1424-8220.