Title :
Adaptive fuzzy system for transform image coding
Author :
Kong, Seong-Gon ; Kosko, Bart
Author_Institution :
Dept. of Electr. Eng., Univ. of Southern California, Los Angeles, CA, USA
Abstract :
An adaptive fuzzy associative memory (AFAM) system is described. It can efficiently classify subimages in adaptive transform image coding. The AFAM system, trained with differential competitive learning for product-space clustering, demonstrated good compressed-image quality at a less than 1-bit-per-pixel rate. It achieved 16-to-1 image compression with only five fuzzy rules. The AFAM system encodes different images without modification and reduces side information when multiple images are encoded. The bank of fuzzy rules estimates the sampled transform-coding process without a mathematical model of how outputs depend on inputs, without mathematical transform techniques, and without rules articulated by experts. The AFAM system provides modular model-free estimation of the transform-coding process
Keywords :
content-addressable storage; data compression; fuzzy logic; neural nets; adaptive fuzzy associative memory; adaptive transform image coding; compressed-image quality; differential competitive learning; fuzzy rules; image compression; mathematical transform techniques; model-free estimation; product-space clustering; transform image coding; Adaptive systems; Decoding; Discrete cosine transforms; Energy measurement; Fuzzy systems; Image coding; Image processing; Pixel; Signal processing; Transform coding;
Conference_Titel :
Neural Networks, 1991., IJCNN-91-Seattle International Joint Conference on
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-0164-1
DOI :
10.1109/IJCNN.1991.155248
