Flexible human-definable automatic behavior analysis for suspicious activity detection in surveillance cameras to protect critical infrastructures

Suspicious activity detection from surveillance video is the main objective of the work presented in this paper. The method developed consist of various stages of suspicious frame detection, and verifying the frame for suspicious activity related analysis of human movements within obtained set of suspicious frames. The method consist of GLCM feature extraction which constitutes the features such as energy, prominence, contrast, entropy, homogeneity type of features and matching using Euclidian distance along with descriptor features obtained by using Harris corner features and cosine similarity index estimation. The successful suspicious activity detection rate is analyzed which shows better performance and time saving method while analyzing large surveillance video dataset.

Download Full-text

Suspicious Activity Detection in Surveillance Footage

2019 International Conference on Electrical and Computing Technologies and Applications (ICECTA) ◽

10.1109/icecta48151.2019.8959600 ◽

2019 ◽

Author(s):

Sathyajit Loganathan ◽

Gayashan Kariyawasam ◽

Prasanna Sumathipala

Keyword(s):

Activity Detection ◽

Suspicious Activity

Download Full-text

Incremental behavior modeling and suspicious activity detection

Pattern Recognition ◽

10.1016/j.patcog.2012.10.008 ◽

2013 ◽

Vol 46 (3) ◽

pp. 671-680 ◽

Cited By ~ 20

Author(s):

Kan Ouivirach ◽

Shashi Gharti ◽

Matthew N. Dailey

Keyword(s):

Behavior Modeling ◽

Activity Detection ◽

Suspicious Activity

Download Full-text

SUSPICIOUS ACTIVITY DETECTION USING CCTV SURVEILLANCE VIDEO

Journal of Information System and Technology Management ◽

10.35631/jistm.622006 ◽

2021 ◽

Vol 6 (22) ◽

pp. 60-70

Author(s):

Bushra Yasmeen ◽

Haslina Arshad ◽

Hameedur Rahman

Keyword(s):

Neural Network ◽

Machine Learning ◽

Deep Learning ◽

Training Model ◽

Activity Detection ◽

Surveillance Video ◽

Full Support ◽

Cctv Surveillance ◽

Developed World ◽

Suspicious Activity

Security has recently been given the highest priority with the rise in the number of antisocial activations taking place. To continuously track individuals and their interactions, CCTVs have been built in several ways. Every person is recorded on an image on average 30 times a day in a developed world with a community of 1.6 billion. The resolution of 710*570 captured at knitting will approximate 20 GB per day. Constant monitoring of human data makes it hard to judge whether the incident is an irregular one, and it is an almost uphill struggle when a population and its full support are needed. In this paper, we make a system for the detection of suspicious activity using CCTV surveillance video. There seems to be a need to demonstrate in which frame the behavior is located as well as which section of it allows the faster judgment of the suspicious activity is unusual. This is done by converting the video into frames and analyzing the persons and their activates from the processed frames. We have accepted wide support from Machine learning and Deep Learning Algorithms to make it possible. To automate that process, first, we need to build a training model using a large number of images (all possible images which describe features of suspicious activities) and a “Convolution Neural Network‟ using the Tensor Flow Python module. We can then upload any video into the application, and it will extract frames from the uploaded video and then that frame will be applied on a training model to predict its class such as suspicious or normal.

Download Full-text