Centre for Robotics Research at the Nitte Meenakshi Institute of Technology, India

Biography: Dr. Jharna Majumdar is currently the Dean R & D, Prof. Dept of M Tech Comp Sc & Engg. and Head, Centre for Robotics Research at the Nitte Meenakshi Institute of Technology, Bangalore, India. Dr. Majumdar served Defense Research and Development Organization, Govt. of India from 1990 to 2007 as Senior Scientist ‘G’ and Head of Aerial Image Exploitation Division, Aeronautical Development Establishment, Bangalore, India. Dr. Majumdar worked as a Research Scientist on ‘Robotics and Automation’ at the Forschungs Zentrum Informatik, Karlsruhe, Germany from 1983 to 1989 and as a Research Scientist at the Stanford Research International California, USA in 2002. Dr. Majumdar has a PhD in Electrical Engg., Undergraduate and Post Graduate Engineering from IIT Kharagpur, India.

Dr. Majumdar published more than 180 reviewed technical papers, has 4 Patents and received a large number of awards for her research. Some of her awards are : Award from President, Stanford Research International (SRI International), California, USA in 2002, Performance Excellence Award from the Prime Minister of India 2004, Dr V M Ghatage award from Aeronautical Society of India 2005, Dr. Suman Sharma Award from National Design and Research Forum (NDRF) 2006, Dr. Kalpana Chawla Memorial Lecture Award 2007, IEEE Award 2011 and CESEM award from the Vision Group on Science and Technology, Bangalore, India in 2016. She is a Fellow of the Aeronautical Society of India, Fellow of the Institution of Engineers and Life Member of the Computer Society of India.

Her project with the Indian Space Research Organization (ISRO) and the students from 7 Engineering Colleges to build the first Smallest Satellite in India (STUDSAT) of PICO Category (weighing 850 Gram) has taken off to the orbit successfully on July 10 2010. Nitte Amateur Satellite Tracking Centre (NASTRAC) developed by a team of students under her guidance is a patent and the First Tracking Station of Small Satellites in the country.

Her current research areas include Real Time Image and Video Processing, Data and Video Analytics, Robotics and Autonomous Systems, Artificial Intelligence, Machine Learning, ANN and Deep Learning.

Speech Title: Advance algorithms of Video Processing on embedded hardware board For Video Content Analysis and real time application in Robotics

Abstract: With the rapid growth of advance technology, availability of affordable low-cost sensors coupled to high end computers and embedded hardware boards made video analysis applicable for real time implementation of critical tasks.

The first part of the talk starts with Video Content Analysis which consists of video shot detection and summarization. Research results show the use of a variety of features such as optical flow with corner detectors, feature extracted from texture and higher order local derivatives to detect the Abrupt Transition ‘CUT’ in Video. The results of shot Transition are used to extract the key-frames using clustering algorithms. We have also shown a new research work on finding ‘Cut’ transition of video in various Transformation Domain, use of appropriate memory model and embedded hardware boards to reduce time intensive high computation cost.

Very few research are reported to determine the Gradual Transition of video viz., Fade, Wipe and Dissolve. In our research work, we have shown the modelling of soft transitions of video using texture as a cue and the principles of machine learning.

Automatic detection and tracking of interesting targets from video is an interesting area of research starting from autonomous navigation of mobile robot, surveillance and security application, Human activity recognition, event detection and many others. In this part of the talk, some of our research results at the Center for Robotics Research will be demonstrated

Keywords: Video Shot Detection, Video Summarization, Abrupt Transition, Gradual Transition, Optical Flow, Texture, Higher Order Derivatives, Human Tracking, Mobile Robot, Embedded Hardware Boards, ANN, CNN