DocumentCode :
3119075
Title :
Variable scale edge detection of images
Author :
Mathur, Bimal P. ; Wang, H.T. ; Tsen, C.T. ; Walton, E.
Author_Institution :
Rockwell Int. Sci. Center, Thousand Oaks, CA, USA
fYear :
1988
fDate :
16-19 May 1988
Abstract :
A set of analog CMOS chips has been designed to perform two-dimensional convolutions between images and Gaussian functions. This implementation, which relies on the transient response of a RC plane, differs from neural net-like implementations proposed by C. Koch et al. (1986) which rely on the steady-state response of resistive networks. However, in common with those and other two-dimensional processors, I/O limitations can only be fully addressed with higher levels of integration, e.g. by using the approach developed at MIT
Keywords :
CMOS integrated circuits; computerised pattern recognition; computerised picture processing; image sensors; Gaussian functions; MIT; RC plane; analog CMOS chips; computerised pattern recognition; image detection; transient response; two-dimensional convolutions; variable scale edge detection; Clocks; Convolution; Image edge detection; Image sampling; Intelligent robots; Land vehicles; Laplace equations; Real time systems; Transient response; Voltage;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Custom Integrated Circuits Conference, 1988., Proceedings of the IEEE 1988
Conference_Location :
Rochester, NY
Type :
conf
DOI :
10.1109/CICC.1988.20853
Filename :
20853
Link To Document :
بازگشت