Dr G Athithan

dr-g-athithanDr. G Athithan, Distinguished Scientist has been appointed as Director General – Micro Electronic Devices & Computational Systems (MED & CoS) wef 01 June 2016. He was born on 30th October 1958. He received his B.E. (Hons) degree in Electronics and Communications in 1981 from Madras University. He received his Ph.D. degree in Physics (Neural Networks) in 1997 from the Indian Institute of Technology, Bombay. After receiving his B.E., he was a trainee in Bhabha Atomic Research Centre, Trombay for a year. In August 1982 he joined the Indira Gandhi Centre for Atomic Research (IGCAR), Kalpakkam. He worked in IGCAR for six years on computer graphics, computer aided design, and modeling of Penrose patterns and crystal structures.
In October 1988 he joined the Advanced Numerical Research and Analysis Group (ANURAG), Hyderabad, where he continued his work on computer graphics besides taking up projects on data visualization, parallel processing, and neural networks. His R&D efforts in ANURAG contributed to the development of ANAMICA and other software for medical and scientific data visualization and also to the design and productionisation of PACE series of parallel computers. As project coordinator to design and implement an Intranet for DRDO, he contributed to the setting up of the first version of DRONA network. During this tenure at ANURAG, he obtained his Ph.D. degree for his work on solving two fundamental problems of the Hopfield model of human memory.
From June 2000 onwards he was with the Centre for Artificial Intelligence and Robotics, Bangalore working in the field of Information Security among others. As the project director of a major initiative to develop Information Security Technologies and Solutions, he brought out a suite of products for preventive and reactive security requirements of the customers of DRDO. In particular, he developed the concept of information security overlay and applied it as a successful model for protection of two tactical C3I networks of the Indian Army. Since Jan 2013 he is with the Scientific Analysis Group, Delhi and has made contributions to the framing of the DRDO Information Security Procedures and guidelines as well as to the drafting of the Cyber Security part of the Long Term Technology Perspective Plan.
His current research interests are Information Security Management, Computational Intelligence, Network Data Mining, and Cyber Security Technologies. So far he has guided two scientists towards their Ph.D. and three scientists towards their M.S. (Engg.) research degrees in reputed institutions of the country. He has participated in a number of international and national conferences and workshops, and has published about twenty-five papers in archived journals and about thirty papers in conference proceedings.
Dr. Athithan was selected as a Young Associate of the Indian Academy of Sciences in 1989. He received the Scientist of the Year award from DRDO for the year 1998. He received the R & D Excellence award for year 2006 from Bharat Electronics Ltd. During 2005 he became a senior member of the IEEE. As the leader of his team, he received the Agni Award for Excellence in Self-Reliance for the year 2008 from DRDO. He was promoted to the rank of Distinguished Scientist in September 2015.

 

Keynote at ADCOM 2018

How do you solve a problem like computer vision?
The sensory world of human experience is dominated by vision involving imagery and video data. Among the four types of media data that are handled by information processing systems too, images and video play an increasingly dominant role besides speech and text. The applications arising out of image and video processing touch almost all aspects of our personal and professional activities. Such applications are vital for business automation as well as for defence forces. A challenging objective in the development of these applications is to make them perform their tasks in a manner that compares favourably with that of the average human brain. Central to the realization of this objective is the solution to the problem of computer vision.
In the lifecycle of an image or a video clip, the starting point is recording. While the earliest methods were passive and sufficed the requirement of a bygone era, today we can computerise the capturing process enabling it with intelligence of the kind that the human eye-brain system effortlessly brings into play. As a result, additional information at pixel level such as depth and motion can be gleaned at the recording stage itself so that segmentation and object recognition are made a lot more feasible. Post recording, if the image and video data were captured in a noisy or misconfigured environment, powerful restoration techniques can be employed to enhance the quality of these poorly captured images and videos. Finally the third and vital stage of processing involves recognizing the classes and identities of the objects and activities present in the images and videos. Mimicking the capability of the human brain in this stage is the most challenging, but the increasing capacities of computing power and storage in modern multi-core and multi-processor servers offer great platforms to tackle this challenge. The talk would focus on the evolutionary changes in the analysis of images and video data over the past few decades. It would highlight the current gaps between computer vision and human vision and outline a memory-intensive method of solving the problem of computer vision.

We value your feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.