Abstract
The image-processing techniques required for an image-understanding system are considered. Although many of the image-processing steps can be implemented with neural nets, it is found that only the classification stage is currently optimally implemented in this way. Conventional methods of image processing that can be used in conjunction with neural net classifiers are described. A case study employing these concepts is presented. Optical realizations of many operations are noted.

This publication has 15 references indexed in Scilit: