Mysore/Mysuru: GSSS Institute of Engineering and Technology for Women (GSSSIETW), Mysuru, IEEE Student Branch and IEEE WIE Affinity Group, in association with IEEE, Bangalore Section and IEEE Circuits and Systems (CAS) Society, Bangalore Chapter, had organised a Webinar on ‘Deep Learning Based Video Analytics for Surveillance IoT Applications’ recently by Dr. Supavadee Aramvith, Associate Professor, Department of Electrical Engineering, Chulalongkorn University, Thailand and Candidate, 2021-2022 IEEE Region 10 Director-Elect.

The session commenced by welcoming the panel members, resource persons and participants by Dr. G. Manjula, Associate Professor, Department of TCE. Webinar was attended by Convener Dr. B.D. Parameshachari, Professor and Head, GSSSIETW, IEEE Student Branch Counsellor, GSSSIETW. The programme was coordinated by M. Keerthi Kumar, Asst. Professor, Dept. of TCE,                                                          GSSSIETW, Mysuru.

Dr. Supavadee Aramvith initiated the Webinar session on ‘Deep Learning Based Video Analytics for Surveillance IoT Applications.’ In her talk, she expressed that, main motivation for choosing video surveillance was to enhance public safety, and also discussed about the major challenges faced to arrive at the solutions.

She gave information about the distinct features of video analytics which includes person detection and tracking, Image Super Resolution, Human feature Extraction, Face Cataloging, Monitoring and warning, Person tracking through multiple co-operative Cameras. She spoke about the research opportunities in surveillance of video Analytics. She explained with an example to train and test machine learning model for Video Analytics Surveillance for IoT Applications.

Dr. Supavadee Aramvith discussed about the current trends and researches in                                                 video analytics. 

As surveillance cameras have been widely installed worldwide, although the main purpose of those cameras is for monitoring, most significant task is to be able to analyse video contents and extract useful information. Deep learning based computer vision techniques utilising multi-layer neural network is drastically improving the performance of video analytics to a certain extent.

Several ongoing researches on deep learning based video analytics such as image super resolution, real-time multiple face recognition system, video anomaly detection and several implementations of embedded video analytic system on FPGA and Single Board Computers were discussed. Some possible scenarios of utilising video analytics, IoT for industries were also mentioned.

Dr. M. Shivakumar, Principal, GSSSIETW, Mysuru, gave the concluding remarks for the Webinar. Dr. S. Jalaja, IEEE, CAS Banglaore Chapter, was the panellist for this Webinar series.