A VLSI neuroprocessor for real-time image flow computing

Wai-Chi  Fang*, Bing J. Sheu, Ji Chien Lee

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

A locally connected multi-layer stochastic neural network and its associated VLSI array neuroprocessors have been developed for high-performance image flow computing systems. An extendable VLSI neural chip has been designed with a silicon area of 4.6 × 6.8 mm2 in a MOSIS 2-μm scalable CMOS process. The mixed analog-digital design techniques are utilized to achieve compact and programmable synapses with gain-adjustable neurons and winner-take-all cells for massively parallel neural computation. Hardware annealing through the control of the neurons' gain helps to efficiently search the optimal solutions. Computing of image flow using one 2-μm 72-neuron neural chip can be accelerated by a factor of 187 more than a Sun-4/260 workstation. Real-time image flow processing on industrial images is practical using an extended array of VLSI neural chips. Actual examples on moving trucks are presented.

Original languageEnglish
Title of host publicationProceedings - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing
Editors Anon
PublisherPubl by IEEE
Pages2413-2416
Number of pages4
ISBN (Print)078030033
DOIs
StatePublished - 1 Dec 1991
EventProceedings of the 1991 International Conference on Acoustics, Speech, and Signal Processing - ICASSP 91 - Toronto, Ont, Can
Duration: 14 May 199117 May 1991

Publication series

NameProceedings - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing
Volume4
ISSN (Print)0736-7791

Conference

ConferenceProceedings of the 1991 International Conference on Acoustics, Speech, and Signal Processing - ICASSP 91
CityToronto, Ont, Can
Period14/05/9117/05/91

Fingerprint

Dive into the research topics of 'A VLSI neuroprocessor for real-time image flow computing'. Together they form a unique fingerprint.

Cite this