ISSN (Online) 2581-9429

IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 14, May 2023

Suspicious Activity Detection Using Video

Surveillance
Prof. O. M. Ghag1, Ms. Vijaya A. Pore2, Ms. Mrunal M. Yedlawar3,
Ms. Vaishnavi Pandhare4, Mr. Samir Mulla5
Assistant Professor, Department of Information Technology1
Students, Department of Information Technology2,3,4,5
Sinhgad Academy of Engineering, Pune, Maharashtra, India

Abstract: Suspicious Activity is predicting the part or joint locations of a person from an image or a video.
This project will detect suspicious human Activity from real-time CCTV footage using neural networks.
Human suspicious Activity is one of the key problems that has been studied for several years. It is important
because of the complete number of applications that can benefit from Activity detection. For example, human
pose estimation is used in applications including video surveillance, animal tracking, and behaviour
understanding, sign language detection, advanced human-computer interaction, and marker less motion
capturing. Low-cost depth sensors have limitations like limited to indoor use, and their low resolution and
noisy depth information make it difficult to estimate human poses from depth images and videos. Hence, we
plan to use neural networks to overcome these problems and find solution on it. Suspicious human activity
recognition from surveillance video is an active research of image, video processing and computer vision.
Through the visual surveillance, human activities can be monitored in sensitive areas and public areas such as
bus stations, railway stations, airports, banks, shopping malls, schools and colleges, parking lots, roads, etc.
to prevent terrorism, theft, accidents and illegal parking, vandalism, fighting, chain snatching, crime, and
other suspicious activities. It is very difficult to watch public places continuously, therefore an intelligent video
surveillance is required that can monitor human activities in real-time and categorize them as usual and
unusual activities; and can generate an alert. The research being carried out is on images and videos. Also,
none of the papers published tries to use CNNs to detect suspicious activities with such accuracy.

Keywords: Suspicious, surveillance, vandalism, snatching

I. INTRODUCTION
With increasing crime rates, it becomes a problem if they are not identified in time and necessary precautionary actions
taken. Most urban and metropolitan areas have surveillance systems installed which constantly accumulates data. With the
vast accumulation of surveillance data there are higher chances of suspicious activities to occur. This problem can be
solved by the automation of the video surveillance. The function of the automated system is to give indication in the form
of alarm or any other form when the predefined abnormal activity is happened. The framework of the system consists of
defining suspicious activity, background subtraction, object detection, tracking and classification of activities. The
suspicious activities are defined using Semantic approach which applies the human understanding of the activity. The
content of the paper are as follows- In section II the work done by various researchers in this field are discussed. Section
III describes the system flow and the working of the system. In section IV, the results which are obtained after experiment
are shown. Section V contains the conclusion which is drawn from the results obtained and the future work.

II. LITERATURE SURVEY

Since the past two decades due to the arrival of various information system and technology, there has been a great increase
and development in Surveillance system. There have been drastic changes in surveillance system and the various ways in
which they are implemented. There are various methods such as Motion Detection, Object Detection, Object Tracking,
Concept of Fractal, and various clustering techniques used to achieve utmost accuracy. India Various businesses whether

Copyright to IJARSCT DOI: 10.48175/IJARSCT-10800 65

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 14, May 2023

large scale or small scale has started increasingly using the managerial database to store the numerous accumulated large
amounts of marketing data, so to keep the information sorted in order, yet still there are various losses and shoplifting,
robbery, break-in in the store is a few to name. Various management related tools and policies such as supply chain
management, customer relationship management, demand management, and customer demand management etc. Started
being used to increase the potential of scale and have a proper track. This entire factor aggregated to purpose a better
system to provide targeted surveillance. Keeping track of sales and customer relationships is one thing and keeping track
of people’s or customer’s activity is another. The last decade witnessed the extensive use of surveillance system in various
public spaces using CCTV and drones. After the events like Mumbai Terror Attack from 2011, there has been great
increasing demand for a behaviour surveillance system that guarantees the peoples safety in the public areas. It should also
include public spaces like football grounds, cricket stadiums, music concerts and places where people gather in large
numbers such as malls. Such places lack to have proper surveillance that ensures the safety of the people. This paper
proposes the utilization of the current surveillance system and upgrading it to a point where it can detect suspicious
activity of the people. In this paper the goal is find a new approach to maximize the accuracy of detecting a suspicious
activity rather than using the previously tested methods like SVM, genetic algorithm, continuous action detection of
actions of interest among actions of non-interest, video visual analytics system and posture representation techniques all
the result in predicting the movement of the customer and the network concepts, random forest algorithm, and deep
learning-based fusion system. Human tracking and motion detection, as well as behaviour analysis are widely researched
topics. Human tracking is also one of the aspects that help determine the behaviour of a person. It acts as a comprehensive
framework for tracking a human motion. Using Machine Learning and deep learning algorithms makes it easy for the
actual interaction between human and computers. This system thus proves to be trust worthy and worth putting time in.
Robbery is a global problem. It’s an open social problem. Although a growing number of CCTV cameras are being
installed at public places (such as airports, banks, shopping malls, etc) yet the conventional surveillance systems that rely
on human operators are inefficient in detecting rare anomalous events, such as robbery, in real-time. So, a robust and
efficient automated surveillance system, able to accurately detect any robbery attempt from the CCTV footage, and can
respond effectively (e.g., raise alarm, lock the vault, call police, release tear gas, etc.) to foil the plot is needed. But first a
surveillance system to detect robbery in real time is must. Finally, it is noticed that the current CCTV system requires a
security guard to constantly monitor all the activities. The efficiency of the guard deteriorates after some time due to
which the robbery in the shop might go undetected. TheSome proposed system like Human Detection and Tracking on
Surveillance Video, Human Motion Analyser, Suspicious Behaviour Recognition based on Face Features had a very
limited scope and could not detect ’Shoplifting’, ‘Robbery’ and ‘Break-In’. Other systems such as Three Stream
C3D+LSTM showed very less accuracy hence could not be practically implemented. Therefore, the main motivation of
this paper is to give CCTV cameras the ability to detect suspicious activities such as ‘Shoplifting’, ‘Robbery’, and ‘Break-
In’ in real-time and notifying the shop owners. S. Nagaprasad et al. [8], Ajay S. Ladkat et al. [9], S. L. Bangare et al. [10-
15], K. Gulati et al. [16], P. S. Bangare et al. [17-18], Xu Wu et al. [19], V. Durga Prasad Jasti et al. [20], A. S. Zamani et
al. [21], M. L. Bangare et al. [22] and S. Mall et al. [23] have proposed various research models which were referred here.

III. METHODOLOGY
Video capture, scene pre-processing, feature extraction, events classification, and prediction are all phases of the
architecture. The idea of this paper is to use Computer Vision (CV) which is widely used to extract useful information
from image or video. CV is a subset of Artificial Intelligence (AI) and Machine Learning (ML). Until recently, CV had
very limited capabilities, but after the invention of deep learning and neural network, it has taken a leap further. It is seen
that CV is able of surpass humans in certain areas related to labelling and detecting objects. This paper aims to propose a
system using Convolution Neural Network, which detects suspicious activity of customer and notifies the shop owner.
System can identify robbers as they tend to cover their face with cloth and point a gun at the shop owner. Break-in
generally happens when the shop is closed i.e., at night. Thus, providing a system that determines suspicious activity for
surveillance is must in today’s world and hence this system delivers such services of tracking all such deception and
forgery and thus making a huge revolution in today’s surveillance system. Database Collection The database contains
images and videos collected from Google and categorised into 3 classes Shoplifting, Robbing and Break-In. Images Pre-
Copyright to IJARSCT DOI: 10.48175/IJARSCT-10800 66
www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 14, May 2023

processing. The collected images and videos from Google were of different resolutions and formats. All the images were
resized to 224x224 to standardize them. After that they were converted to RGB format. Then the process of data
augmentation is done to avoid overfitting or underfitting. In data augmentation, rotation, zoom, shear range, width and
height shift, and flip operations were performed on each image. It is then given as input to the Convolutional Neural
Network (CNN). Model Training For training the model transfer learning is used by using pre-trained ImageNet weights,
instead of training the CNN model from scratch. The full architecture is depicted. A model is generated by training the
CNN with epoch. This paper evaluates the performance of the model intwo phases. The first phase consists of evaluating
the performance on the test set of images and videos collected from Google. Considering the accuracy, precision and recall
of the system as the evaluation metric. To show the trade-off between true Operating Characteristics curve. Finally, the
plotted precision-recall and area under curve is used positive and false positive rate plotting.

IV. SYSTEM ARCHITECTURE

The Receiver to analyse the performance of the model. The second phase consists of providing real time videos as input to
the system. The frames are extracted from a live video stream source. The extracted frame is then pre-processed by first
resizing it (the frame is resized to the size of the image provided during the training phase of the model), followed by RGB
channel ordering. Each frame extracted and pre-processed from the real-time video is given to the model, which then
predicts its class label. The label of the class with the highest probability is returned by the model. The same process is
repeated for all the frames from the real-time video. For every input frame, its predicted class along with its probability is
displayed on the output frame. The videos for the input are drawn from both existing and newly developed datasets.
Frames are retrieved from collected videos as part of the pre-processing process. The labelled folders are formed based on
the videos, and the frames are placed in them. The entire video is changed to frames, which are saved as files. Every frame
is then scaled to (224 x 224) pixels to fit inside the 2DCNN architecture and saved. The test video is also saved in a folder
after being converted to frames and scaled to 224 x 224. For video pre-processing, the Python package Open CV is used.
Layers of Convolution, a layer of ReLU (Rectified Linear Unit) activation function, Layers of max pooling, Layers of
fully connected dense, and Layers of normalizing are all represented. The model can be fine-tuned to our specifications,
and the model's last layer is deleted. The model is then trained using LSTM (Long Short-Term Memory) architecture. In
sequence prediction problems, the networks of LSTM are a type of RNN that can learn order dependency. Now there is a
ReLU activation function, a layer of dropout, and layers of dense that are fully coupled. The system categorizes the videos
as either unusual (fainting or fighting) or regular (running, walking). An SMS (Short Message Service) will be sent to the
appropriate officials in the event of suspicious conduct.

Fig.6:Architecture Diagram
Copyright to IJARSCT DOI: 10.48175/IJARSCT-10800 67
www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science,, Communication and Technology (IJARSCT)
International Open-Access,
Access, Double-Blind,
Double Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 14, May 2023

V. SCOPE
Inn our model we take number of images form video if video is large then it will take more time to create frames. In future
we try to improve accuracy and make sure that it takesless
takes time to capture human suspicious activity. In future we will add
more images data set to detect suspicious activity and fight.

VI. RESULT
After clear understanding and execution of the project. It has classified normal and suspicious activities. These were the
Suspicious Activities detected from the uploaded CCTV video Surveillance.

Registration Page

Fig.7: Result of Working Module

Login Page

Fig.8: Result of Working Module

Copyright to IJARSCT DOI: 10.48175/IJARSCT-10800 68

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science,, Communication and Technology (IJARSCT)
International Open-Access,
Access, Double-Blind,
Double Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 14, May 2023

Selecting Video

Fig.9: Result of Working Module

Detecting Activity

Fig.10: Activity Detection

VII. CONCLUSION
Nowadays, practically everyone understands the importance of CCTV footage, yet in most circumstances, this footage is
only utilized for purposes of investigation after an incident or crime has occurred. The proposed system offers the
advantage of sending alerts when an accident occurred. CCTV footage is being outcome is a directive to appropriate
officials to decide if the result shows that an undesirable event is likely to occur. As a resul
result, this can be prevented.
Although the proposed approach is dedicated to the academic’s realm, it can be utilized to forecast more unusual actions in
private or public settings. The system can be utilized in any location where training should be delivered in conjunction
with the unusual activity that is appropriate for that location.

REFERENCES
[1]. S. Zaidi, B. Jagadeesh, K. V. Sudheesh and A. A. Audre, “Video
Video Anomaly Detection and Classification for
Human Activity Recognition,” 2017 International Conference
Conference on Current Trends in Computer, Electrical,
Electronics and Communication (CTCEEC), Mysore, 2017, pp. 544-548.
544
[2]. S. Dodge and L. Karam, “A A Study and Comparison of Human and Deep Learning Recognition Performance
under Visual Distortions,” 2017 26th International Conference on Computer Communication and Networks
(ICCCN), Vancouver, BC, 2017, pp. 1-7.
1

Copyright to IJARSCT DOI: 10.48175/IJARSCT-10800 69

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 14, May 2023

[3]. K. S. Kushwaha, C. M. Sharma, M. Khare, R. K. Srivastava and A. Khare, “Automatic multiple human
detection and tracking for visual surveillance system,” 2012 International Conference on Informatics,
Electronics & Vision (ICIEV), Dhaka, 2012, pp. 326- 331.
[4]. W. Sultani, C. Chen and M. Shah,” Real-World Anomaly Detection in Surveillance Videos,” 2018 IEEE/CVF
Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, 2018, pp. 6479- 6488.
[5]. A. Nurhopipah, A. Harjoko: “Motion Detection and Face Recognition for CCTV Surveillance System”, IJCCS
(Indonesian Journal of Computing and Cybernetics Systems) Vol.12, No.2, July 2018, pp. 107 118, 2018.
[6]. A. Eamthanakul, M. Ketcham and N. Chumuang, “The Traffic Congestion Investigating System by Image
Processing from CCTV Camera,” 2017 International Conference on Digital Arts, Media and Technology
(ICDAMT), Chiang Mai, 2017, pp. 240-245.
[7]. B. A., P. P. and V. S., “Detection of Suspicious Human Activity based on CNN-DBNN Algorithm for Video
Surveillance Applications,” 2019 Innovations in Power and Advanced Computing Technologies (i-PACT),
Vellore, India, 2019, pp. 1-7
[8]. S. Nagaprasad, D. L. Padmaja, YaserQuereshi, S.L. Bangare, Manmohan Mishra, Mazumdar B. D.,
“Investigating the Impact of Machine Learning in Pharmaceutical Industry”, Journal of Pharmaceutical
Research International (Past name: British Journal of Pharmaceutical Research, Past ISSN: 2231-2919, NLM
ID: 101631759), Volume 33, Issue 46A, Pages 6-14, Publisher: JPRI
https://www.journaljpri.com/index.php/JPRI/article/view/32834
[9]. Ajay S. Ladkat, Sunil L. Bangare, Vishal Jagota, Sumaya Sanober, Shehab Mohamed Beram, Kantilal Rane,
Bhupesh Kumar Singh, “Deep Neural Network-Based Novel Mathematical Model for 3D Brain Tumor
Segmentation", Computational Intelligence and Neuroscience, vol. 2022, Article ID 4271711, 8 pages, 2022.
https://doi.org/10.1155/2022/4271711
[10]. S. L. Bangare, “Brain Tumor Detection Using Machine Learning Approach”, Design Engineering ISSN: 0011-
9342,Scopus Index- Q4, EiCompendex, Volume 2021, Issue 7, Pages 7557-7566, Publisher Design
Engineering.
[11]. S. L. Bangare, and P. S. Bangare. "Automated testing in development phase." International Journal of
Engineering Science and Technology 4.2 (2012): 677-680.
[12]. S. L. Bangare, N. B. Dhawas, V. S. Taware, S. K. Dighe, & P. S. Bagmare, (2017). “Implementation of fabric
fault detection system using image processing”, International Journal of Research in Advent Technology, Vol.5,
No.6, June 2017, E-ISSN: 2321-9637.
[13]. S. L. Bangare, N. B. Dhawas, V. S. Taware, S. K. Dighe, & P. S. Bagmare (2017). “Fabric fault detection using
image processing method”, International Journal of Advanced Research in Computer and Communication
Engineering, 6(4), 405-409.
[14]. S. L. Bangare, S., H. Rajankar, P. Patil, K. Nakum, G. Paraskar, (2022). “Pneumonia detection and
classification using CNN and VGG16”. International Journal of Advanced Research in Science,
Communication and Technology, 12, 771-779.
[15]. Sunil L. Bangare, Deepali Virmani, Girija Rani Karetla, Pankaj Chaudhary, Harveen Kaur, Syed Nisar Hussain
Bukhari, Shahajan Miah, "Forecasting the Applied Deep Learning Tools in Enhancing Food Quality for Heart
Related Diseases Effectively: A Study Using Structural Equation Model Analysis", Journal of Food Quality,
vol. 2022, Article ID 6987569, 8 pages, 2022. https://doi.org/10.1155/2022/6987569
[16]. K. Gulati, M. Sharma, S. Eliyas, & Sunil L. Bangare (2021), “Use for graphical user tools in data analytics and
machine learning application”, Turkish Journal of Physiotherapy and Rehabilitation, 32(3), 2651-4451.
[17]. P. S. Bangare, Ashwini Pote, Sunil L. Bangare, Pooja Kurhekar, and Dhanraj Patil, "The online home security
system: ways to protect home from intruders & thefts." International Journal of Innovative Technology and
Exploring Engineering (IJITEE) ISSN (2013): 2278-3075.
[18]. P. S. Bangare, S. L. Bangare, R. U. Yawle and S. T. Patil, "Detection of human feature in abandoned object
with modern security alert system using Android Application," 2017 International Conference on Emerging
Trends & Innovation in ICT (ICEI), Pune, India, 2017, pp. 139-144, doi: 10.1109/ETIICT.2017.7977025.
Copyright to IJARSCT DOI: 10.48175/IJARSCT-10800 70
www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 14, May 2023

[19]. Xu Wu, Dezhi Wei, Bharati P. Vasgi, Ahmed Kareem Oleiwi, Sunil L. Bangare, Evans Asenso, "Research on
Network Security Situational Awareness Based on Crawler Algorithm", Security and Communication
Networks, vol. 2022, Article ID 3639174, 9 pages, 2022. https://doi.org/10.1155/2022/3639174.
[20]. V. Durga Prasad Jasti, Enagandula Prasad, Manish Sawale, ShivlalMewada, Manoj L. Bangare, Pushpa M.
Bangare, Sunil L. Bangare, F. Sammy, "Image Processing and Machine Learning-Based Classification and
Detection of Liver Tumor", BioMed Research International, vol. 2022, Article ID 3398156, 7 pages, 2022.
https://doi.org/10.1155/2022/3398156
[21]. Zamani, A. S., Dr. Seema H. Rajput, Dr. Harjeet Kaur, Dr.Meenakshi, Dr. Sunil L. Bangare, & Samrat Ray.
(2022). Towards Applicability of Information Communication Technologies in Automated Disease Detection.
International Journal of Next-Generation Computing, 13(3). https://doi.org/10.47164/ijngc.v13i3.705.
[22]. M. L. Bangare, P. M. Bangare, R. S. Apare, & S. L. Bangare, (2021). “Fog computing-based security of IoT
application”, Design Engineering, 7, 7542-7549.
[23]. S. Mall, A. Srivastava, B. D. Mazumdar, M. Mishra, S. L. Bangare, & A. Deepak, (2022). “Implementation of
machine learning techniques for disease diagnosis”, Materials Today: Proceedings, 51, 2198-
2201.https://www.sciencedirect.com/science/article/abs/pii/S2214785321072679#!

Copyright to IJARSCT DOI: 10.48175/IJARSCT-10800 71

www.ijarsct.co.in