THE LOEVE-KARHUNEN EXPANSION AS A MEANS OF INFORMATION COMPRESSION FOR CLASSIFICATION OF CONTINUOUS SIGNALS.
Abstract
The present paper is concerned mainly with the aspect of information compression, which is only part of the process of recognition. The problem of zoning (division of the space into disjoint volumes corresponding to classes) and the problem of decision making (such as the Bayesian algorithm) require, among others, careful study in connection with classification and recognition, but are not included in this paper.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1965
- Accession Number
- AD0628684
Entities
People
- Satosi Watanabe
Organizations
- IBM Thomas J. Watson Research Center