A neural net based architecture for the segmentation of mixed gray-level and binary pictures

A neural-net-based architecture is proposed to perform segmentation in real time for mixed gray-level and binary pictures. In this approach, the composite picture is divided into 16*16 pixel blocks, which are identified as character blocks or image blocks on the basis of a dichotomy measure computed by an adaptive 16*16 neural net. For compression purposes, each image block is further divided into 4*4 subblocks and, similar to the classical block truncation coding (BTC) scheme, a one-bit nonparametric quantizer is used to encode 16*16 character and 4*4 image blocks. In this case, however, the binary map and quantizer levels are obtained through a neural net segmentor over each block. The efficiency of the neural segmentation in terms of computational speed, data compression, and quality of the compressed picture is demonstrated. The effect of weight quantization is also discussed. VLSI implementations of such adaptive neural nets in CMOS technology are described and simulated in real time for a maximum block size of 256 pixels. >

[1]  David Best.  Can creativity be taught , 1982 .

[2]  T. Troudet,et al.  An adaptive neural net approach to the segmentation of mixed gray-level and binary pictures , 1988, IEEE 1988 International Conference on Neural Networks.

[3]  Paul Wintz,et al.  Digital image processing (2nd ed.) , 1987 .

[4]  Z. Czarnul.  Design of voltage-controlled linear transconductance elements with a method pair of FET transistors , 1986 .

[5]  Thomas S. Huang.  Coding of Two-Tone Images , 1977, IEEE Trans. Commun..

[6]  J. J. Hopfield,et al.  “Neural” computation of decisions in optimization problems , 1985, Biological Cybernetics.

[7]  Yannis Tsividis,et al.  Floating voltage-controlled resistors in CMOS technology , 1982 .

[8]  Stephen M. Walters,et al.  Neural network architecture for crossbar switch control , 1991 .

[9]  Robert B. Allen,et al.  Stochastic Learning Networks and their Electronic Implementation , 1987, NIPS.

[10]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[11]  Jorma Rissanen,et al.  Compression of Black-White Images with Arithmetic Coding , 1981, IEEE Trans. Commun..

[12]  H. Gharavi,et al.  CCITT compatible coding of multilevel pictures , 1983, The Bell System Technical Journal.

[13]  Y. Yasuda,et al.  Data compression for check processing machines , 1980, Proceedings of the IEEE.

[14]  J J Hopfield,et al.  Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[15]  Vishwas Udpikar,et al.  BTC Image Coding Using Vector Quantization , 1987, IEEE Trans. Commun..

[16]  R. Hunter,et al.  International digital facsimile coding standards , 1980, Proceedings of the IEEE.

[17]  J J Hopfield,et al.  Neurons with graded response have collective computational properties like those of two-state neurons. , 1984, Proceedings of the National Academy of Sciences of the United States of America.

[18]  D. Frohman-Bentchkowsky.  Famos—A new semiconductor charge storage device , 1974 .

[19]  Paul Losleben,et al.  Advanced Research in VLSI , 1987 .