Title :
Decoupled-voting Hamming associative memory networks
Author :
Watta, Paul B. ; Akkal, Mohammed ; Hassoun, Mohamad H.
Author_Institution :
Dept. of Electr. & Comput. Eng., Wayne State Univ., Detroit, MI, USA
Abstract :
A class of high performance associative memory models are developed by generalizing the operation of the Hamming associative memory. Two models are proposed: the decoupled Hamming memory and the two-level decoupled Hamming memory. By utilizing local Hamming distance computations, these models are more amenable to practical hardware implementation on parallel machines and available accelerator boards. Simulation results indicate that the proposed associative memory models offer substantial error correction capability
Keywords :
associative processing; content-addressable storage; image recognition; neural nets; Hamming associative memory; Hamming distance; decoupled Hamming memory; decoupled-voting; error correction; image processing; image recognition; Associative memory; Hamming distance; Hardware; Impedance matching; Input variables; Pattern matching; Pattern recognition; Stationary state; Topology;
Conference_Titel :
Neural Networks,1997., International Conference on
Conference_Location :
Houston, TX
Print_ISBN :
0-7803-4122-8
DOI :
10.1109/ICNN.1997.616201