Kimihiro NISHIO^*, Hiroo YONEZU¹, and Yuzo FURUKAWA¹

Department of Electrical and Computer Engineering, Yonago National College of Technology, 4448 Hikona-cho, Yonago, Tottori 683-8502, Japan
¹Department of Electrical and Electronic Engineering, Toyohashi University of Technology, 1-1 Hibarigaoka, Tempaku-cho, Toyohashi, Aichi 441-8580, Japan

(Received November 20, 2006; Accepted May 21, 2007)

We proposed in this research a novel two-dimensional network based on the frog visual system, with a motion detection function and a newly developed simple-shape recognition function, for use in object discrimination by integrated circuits. Specifically, the network mimics the signal processing of the small-field cell in a frog brain, consisting of the tectum and thalamus, which generates signals of the motion and simple shape of an object. The proposed network is constructed from simple analog complementary metal oxide semiconductor (CMOS) circuits; a test chip of the proposed network was fabricated with a 1.2 μm CMOS process. Measurements on the chip clarified that the proposed network can generate signals of the moving direction, velocity, and simple shape, as well as perform information processing of the small-field cell. Results with the simulation program with integrated circuit emphasis (SPICE) showed that the analog circuits used in the network have low power consumption. Applications of the proposed network are expected to realize advanced vision chips with functions such as object discrimination and target tracking.

Key words: analog integrated circuit, motion detection, shape recognition, small-field cell, vision chip

^*E-mail address: nishio@yonago-k.ac.jp