Sanjit K. Mitra is the 2005 SPIE Technology Achievement Award recipient not for a specific innovation, per se, but rather for a lifetime achievement of innovation. Mitra's accomplishments in the fields of signal, image, and video processing are immense, and the number of awards and honors he has received from institutes and organizations all over the world speak to this fact.
Two of the most prestigious awards he has received are the Society Award of the IEEE Signal Processing Society in 2001 and the 2001 Technical Achievement Award from the European Association for Signal Processingboth of which are the highest honor given by those societies. In addition, he is a Fellow of SPIE, the IEEE, and the American Association for the Advancement of Science.
"Sanjit Mitra has helped in shaping the field of image processing from a peripheral area in electrical and computer engineering to one of the most vibrant and core areas," says Ramesh Jain of Georgia Institute of Technology's College of Computing (Atlanta, GA). Career Evolution
After earning his BSc and MSc degrees from Utkal University and the University of Calcutta, respectively, Mitra was an assistant engineer at the Indian Statistical Institute in Calcutta for two years during the late 1950s. It was there that he worked on the first electronic computer in Indiathe Hollerith Electronic Computer, HEC-2M.
Left to right: Alessandro Neri (University of Rome Tre), doctoral student Serkan Hatipoglu, Sanjit Mitra, and Marco Carli (Universtiy of Rome Tre) at Mitra's lab in July 2001.
"It was a big machine with 1000 drum memory, 16-bit, that you had to program in machine language," Mitra says. "We didn't have any [computer] courses, but we learned the hard way how to repair it, and it was down most of the time."
This work uncovered his love for computing, and he soon moved to the United States to study at the University of California, Berkeley, as a teaching assistant. "The only gentleman who taught computer engineering was on sabbatical, so I took a course on circuits by another professor," Mitra says. "I thought the mathematics were so beautiful, so I went to circuits."
Mitra went on to earn an MS and PhD from UC Berkeley, and for more than 30 years researched and taught in the areas of analog circuits and digital signal processing, first at Cornell University, then at UC Davis, and UC Santa Barbara.
"I was getting bored, . . . and the department chair at the University of California, Santa Barbara, asked if I would like to teach image processing. I found it very fascinating that at last I could see the results of algorithms. In signal processing you tell of algorithms you don't see, but here I could see the effect on images. So I got fascinated and I've been there since then, now mostly into video processing."
His and his students' breakthroughs in image and signal processing make for a long list. Just a few of the highlights are included here. In the area of image filtering, he developed a series of efficient algorithms based on quadratic Volterra filters for edge-preserving image enhancement, zooming, and halftoning, as well as robust texture analysis.
He developed various low-complexity methods to achieve high-compression ratio. In one of these patented approaches using subband coding for still pictures, he achieved results better than the international JPEG standard.
His work on rate-distortion analysis using the number of consecutive zero runs as a parameter has been found very effective in predicting the rate-distortion performance of transform-based video coding systems. In fact, this work earned him and his students the IEEE Transactions on Video Technology Best Paper Award.
"His work has most definitely impacted every aspect of the popular image and video processing approaches including image filtering, image compression, image analysis, and image enhancement," says SPIE Fellow Sethuraman Panchanathan, Arizona State University (Tempe, AZ).
"Currently most of my work is trying to develop a metric for video quality based on psycho-physical experiments," Mitra explains. "A digital video needs to be compressed to save the transmission bandwidth and also to save storage. However, when a compressed video is decompressed, you see various artifacts depending on the degree of compression. Industry is interested in finding out how much they can compress before consumers don't like it."
Mitra has been working with Professor John Foley in UC Santa Barbara's Department of Psychology for the past six years to try to understand and quantify this perceptual quality of a video.
"It will be very difficult to come up with an exact number because it's content dependant. The other thing which I also worry about . . . is the effect of audio on the perceptual quality of a video," Mitra says. They plan to conduct psycho-physical experiments to determine the effect of both audio and color on the human perception of video quality. World Traveler
One unique aspect of Mitra's career is the amount he has traveled as part of his research. He has served as visiting professor nearly 20 times in a dozen different countries from Finland to Brazil to Japan.
Mitra is now working on a side project organizing international conferences on signal processing in countries where travel abroad is typically prohibitive. Last year, he organized such a conference in Tunisia. He's currently putting together conferences in Morocco, and is assisting with another in Romania. His main aim is to bring people from other countries to these places to attend the event and network.
"It is fascinating to work with people who have different perspectives, different ways of thinking," he says. "I enjoy meeting people, knowing their culture, and it's been extremely satisfying to me."