Accession Number : AD0708510

Title :   OPTIMUM TECHNIQUES FOR LINEAR FEATURE EXTRACTION.

Descriptive Note : Technical rept.,

Corporate Author : TEXAS UNIV AUSTIN ELECTRONICS RESEARCH CENTER

Personal Author(s) : Henderson,T. L. ; Lainiotis,D. G.

Report Date : 01 DEC 1969

Pagination or Media Count : 144

Abstract : Methods are developed for processing a set of measurements to extract their most important features. The problem is approached from the viewpoint of pattern recognition, assuming two pattern classes, with particular emphasis given to the signal-in-noise detection problem. It is supposed that the feature extractor transforms the measurements into a finite-dimensional feature space, of smaller dimension than the measurement space. (Author)

Descriptors :   (*INFORMATION THEORY, *PATTERN RECOGNITION), CURVE FITTING, DIFFERENCE EQUATIONS, ALGORITHMS, STOCHASTIC PROCESSES, RANDOM VARIABLES, SIGNALS, DATA PROCESSING, OPTIMIZATION, STATISTICAL ANALYSIS

Subject Categories : Cybernetics
      Bionics

Distribution Statement : APPROVED FOR PUBLIC RELEASE