Feature Set Evaluation for Classifiers.

摘要

The activity proposed meet the following five technical objectives: (1) identify and propose a collection of feature set evaluation algorithmic tools which address unique characteristics of feature sets used in ATR applications. (2) implement at least one new promising feature set evaluation algorithm in FORTRAN and integrate it within the On-Line Pattern Analysis and Recognition System (OLPARS), which is an existing commercial software system which provides general purpose feature set evaluation and classifier design capabilities. (3) demonstrate the performance of the new feature set evaluation algorithms already within OLPARS using feature sets derived from both real and simulated E/O imagery. (4) provide DoD with a licensed VAX-compatible copy of the augmented OLPARS software package. (5) document the proposed new set of feature set evaluation algorithms and the test results obtained with the newly implemented algorithm within a final technical report. (jhd)