Image and video processingMLAIComputational neuroscience
I have always been baffled since childhood, about how our human visual system works and how the brain perceives what we see. Colours fascinated me. I was amazed by the way certain colours calmed people while others plainly agitated a human being. The eye was, according to me the best part in God’s creations. I was puzzled by how accurately the human eye, the size of a golf ball with few layers of tissues and flesh, interprets frequencies and colours and has capabilities way beyond of any visual device built by man till date. These ideas about nature and philosophy, I believe, have subconsciously got me interested in signal processing and neural networks.
I did my undergraduate course in electrical and electronic engineering. During the four year duration of this course, I gained in-depth understanding of the various techniques involved in problem solving with mathematical tools and the fundamentals necessary to venture on my own into many fields outside the lab. It was at this point that I started exploring my primary interest in visual systems and the technical aspects of the human brain. I had to teach my self the basics of neural networks because the course was not offered in my college. As I learned more, I started applying my knowledge practically. My final year project was an automatic scanning turret that recognizes targets using neural networks and automatically fires a laser beam at them using monoscopic distance approximation. It also secured the highest score in the entire department. This was coded in C++ on a Linux platform.
After my graduation, I felt that while I had educated myself a little on my interests, it would be greatly beneficial if I were to work on a research project with an experienced guide for awhile. I approached Dr. Srinivasa Chakravarthy of IIT Madras (School of Biosciences) regarding this. On inspecting my profile, he took me in as a research assistant in the computational neuroscience lab. Being in IITM would easily qualify for a life’s achievement for many here but for me it was just the beginning. The first three months of my research study equipped me with the theoretical aspects of neural networks and coding in MatLab and C++. Following this, I worked on a project for hand writing character recognition of Indian languages and its application in the areas like census data collection, encumbrance certificate and accident report. During this period, Prof. Srinivasa Chakravarthy suggested that I parallelly work on other applications of Image processing as well. He referred me to Prof. Krishnan Balasubramaniam from the Mechanical department, who is also the head of the Centre for Non-Destructive Evaluation at our Institute. I worked on an ADR (Automatic Defect Recognition) program for Hancock valves for BHEL (Bharat Heavy Electric Limited, India). The project was successfully completed and a patent has also been filed for the same. Recognizing my efforts in this project, BHEL has engaged me to implement this technology for Automatic defect recognition in pipes and other hardware too. I have published a paper on NDE ADR of Hancock Valve Welds in the Journal of Non-Destructive Testing & Evaluation 2010. Apart from regular project activities, I have attended workshops and seminars, the prominent among them being the winter school for audio and speech processing conducted at IIT, Madras.