Title :
Unconstrained handwritten digit VLSI recognition system based on combined neural networks
Author :
Li, Guoxing ; Shi, Bingxue
Author_Institution :
Dept. of Microelectron., Tsinghua Univ., Beijing, China
Abstract :
A totally unconstrained digit recognition system based on cellular neural network (CNN) and multilayer perceptron (MLP) and its analog-digital mixed mode VLSI implementation is presented in this paper. The CNN is used to extract CCD (connected component detector) feature from 24×24 normalized digit image in horizontal, vertical directions and two diagonal lines row by row. These features are fed into the two layers MLP in time sharing mode after proper compression. MLP has 60×20×10 structure with local connection. It consists of 10×10 neural processing unit (NPU) array which contains latches, weight generating circuits and computation circuits, switched current integrators, current threshold unit and control logic block. The weights are programmable and their control codes come from off-chip EPROM. This recognition system is very smart and effective, it can be implemented in standard digit CMOS technology
Keywords :
EPROM; VLSI; cellular neural nets; character recognition equipment; feature extraction; flip-flops; handwritten character recognition; integrating circuits; mixed analogue-digital integrated circuits; multilayer perceptrons; neural chips; CCD feature extractors; analog-digital mixed mode VLSI implementation; backpropagation; cellular neural network; combined neural networks; computation circuits; connected component detector feature; control codes; control logic block; current threshold unit; handwritten digit VLSI recognition system; latches; multilayer perceptron; nonlinear transfer function; off-chip EPROM; programmable weights; smart system; standard digit CMOS technology; switched current integrators; time sharing mode; unconstrained digit recognition system; weight generating circuits; Analog-digital conversion; CMOS technology; Cellular neural networks; Charge coupled devices; Circuits; Detectors; Handwriting recognition; Logic arrays; Multilayer perceptrons; Very large scale integration;
Conference_Titel :
Solid-State and Integrated Circuit Technology, 1998. Proceedings. 1998 5th International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-4306-9
DOI :
10.1109/ICSICT.1998.785893
