Title :
Neural network representation and implementation of gray scale morphological operators
Author :
Ko, Sung-Jea ; Morales, Aldo
Author_Institution :
Dept. of Electr. & Comput. Eng., Michigan Univ., Dearborn, MI, USA
Abstract :
A neural network implementation of gray-scale operators is introduced. It is based on fuzzy set theory. In this structure, synaptic weights are represented by a gray-scale structuring element. Two learning algorithms are used to train the networks. The first algorithm utilizes the overall equality index. The second algorithm is based on the averaged least-mean square (LMS). It is shown that the LMS-based algorithm is simpler and more robust
Keywords :
filtering and prediction theory; fuzzy set theory; image recognition; learning (artificial intelligence); least squares approximations; neural nets; averaged least-mean square; fuzzy set theory; gray scale morphological operators; gray-scale structuring element; learning algorithms; neural network implementation; overall equality index; synaptic weights; Artificial neural networks; Computer networks; Concurrent computing; Educational institutions; Fuzzy neural networks; Fuzzy sets; Least squares approximation; Morphology; Neural networks; Shape measurement;
Conference_Titel :
Circuits and Systems, 1992. ISCAS '92. Proceedings., 1992 IEEE International Symposium on
Conference_Location :
San Diego, CA
Print_ISBN :
0-7803-0593-0
DOI :
10.1109/ISCAS.1992.230003
