RESEARCH ON DYNAMIC SELF-ORGANIZING PATTERN RECOGNITION MODEL.
Abstract
The report presents the results of an investigation into the properties of a dynamic pattern-recognition system involving a scanner whose operation was suggested by some of the properties of the human eye. The output from the scanner is a sequence of the primitive percepts into which the pattern may be decomposed, and the sequence itself constitutes a sample function of a Markov process which is unique for each pattern class. Various decision processes are suggested for identifying the process, and, hence, the pattern. Neural analog networks are used to identify the primitive percepts. A learning procedure is presented for identifying patterns for which a list of attributes has been measured. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1962
- Accession Number
- AD0690587
Entities
People
- Robert R. Evans