Masatoshi Ishikawa

High-speed Parallel Processing Vision Chip and Its Applications in Dynamic Intelligent Systems

Abstract: High-speed image processing can be implemented by using on-sensor parallel processing vision chips and opens new applications in various fields such as robotics, factory automation, human interface, bio/medical applications, 3D achieving, autonomous vehicles, and so on.

We developed a pixel wise fully parallel processing system (64x64 pixels) as a scale-up model of vision chip in 1993 and an integrated fully parallel processing vision chip (64x64) in 1999. In 2017, our laboratory and Sony developed a stacked CMOS vision chip which has a layer of a high-speed CMOS imager (1,296x976 pixels) and a layer of a column parallel processing element array (1,304 PEs). Those chips can carry out high-speed image processing at 1,000 fps with low latency in the chips.

As the video rate is not enough for operating moving or deforming objects in physical world and controlling dynamics of intelligent robots, so high-speed image processing is key technology for realizing high-speed and low latency intelligent systems based on visual feedback.

In this talk, devices technology, processing architectures, and application systems such as high-speed intelligent robots, dynamic projection mapping, and 1ms 3D shape measurement will be explained.

Masatoshi ISHIKAWA received the B.E., M.E. and Dr. Eng. degrees in mathematical engineering and information physics in 1977, 1979 and 1988, respectively, from the University of Tokyo, Tokyo, Japan. He was a researcher at Industrial Products Research Institute, Tsukuba, Japan, from 1979 to 1989. He moved to the University of Tokyo as an associate professor in 1989. He was a professor of mathematical engineering and information physics, and information physics and computing at University of Tokyo from 1999 to 2001 and from 2001 to 2005, respectively. He was an executive adviser to the president, a vice-president and an executive vice-president of the University of Tokyo, from 2004 to 2005, from 2004 to 2005, and from 2005 to 2006, respectively. He is a project professor at the Data Science Division, Information Technology Center, University of Tokyo from 2020. His current research interests include high-speed massively parallel vision, sensor fusion, high-speed intelligent robots, visual feedback, dynamic intelligent systems, and dynamic interaction.