Abstract
Multilayer Hopfield and Hamming neural Networks structures are proposed for processing and recognition of non binary input patterns. Proposed structures are based on splitting the input pattern into N binary input patterns, where N is the number of bits used for representing each pixel of the original input pattern. Subsequently each binary pattern is processed for a Hopfield (Hamming) neural network. Finally the outputs of each binary neural network are used to reconstruct or identify the non binary input pattern. Simulation results show that proposed structures performs fairly well for input patterns with up to 50% of their pixels distorted.
Original language | English |
---|---|
Pages (from-to) | 775-779 |
Number of pages | 5 |
Journal | National Conference Publication - Institution of Engineers, Australia |
Volume | 2 |
Issue number | 94 /9 |
State | Published - 1994 |
Externally published | Yes |
Event | Proceedings of the International Symposium on Information Theory & Its Applications 1994. Part 1 (of 2) - Sydney, Aust Duration: 20 Nov 1994 → 24 Nov 1994 |