Optical Neural Nets for Scene Analysis
Abstract
Our approach is hybrid and multidisciplinary. We marry pattern recognition and neural net techniques. We also marry optical and digital technologies. A hybrid neural net thus results. We also concentrate on the use of one basic hybrid architecture that is useful for implementing various optimization and adaptive neural nets. Our work thus distinguishes between these two general classes of neural nets (optimization and adaptive) with both being realizable on the same basic hybrid architecture. (KR)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 23, 1990
- Accession Number
- ADA229794
Entities
People
- David P. Casasent
Organizations
- Carnegie Mellon University