Share Email Print

Proceedings Paper

Image processing based on supervised learning in neural networks
Author(s): Akira Hasegawa; Wei Zhang; Kazuyoshi Itoh; Yoshiki Ichioka
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

We applied a learning network to the image processing of human cornea endothelial micrograms. Our neural model is a feed-forward network with interconnections between units constrained to be locally space-invariant to perform space-invariant processing. The network is trained by the error back propagation with some small parts of micrograms and their cell membrane images which are outline drawing made by hand. After training, the network showed good performance with unexpenenced micrograms. The final membrane images were obtained after additional processing by a conventional digital filter that is based on mathematical morphology and linear filtering.

Paper Details

Date Published: 1 November 1991
PDF: 6 pages
Proc. SPIE 1621, Optical Memory and Neural Networks, (1 November 1991); doi: 10.1117/12.50444
Show Author Affiliations
Akira Hasegawa, Osaka Univ. (Japan)
Wei Zhang, Osaka Univ. (Japan)
Kazuyoshi Itoh, Osaka Univ. (Japan)
Yoshiki Ichioka, Osaka Univ. (Japan)

Published in SPIE Proceedings Vol. 1621:
Optical Memory and Neural Networks
Andrei L. Mikaelian, Editor(s)

© SPIE. Terms of Use
Back to Top