Regular Correspondence: Classification Quality Assessment for a Generalized Model-Based Object Identification System

Abstract

An object recognition system, based on global features and nearest neighbor matching, is extended and enhanced using classification quality assessment (CQA) methodology. For quality assessment, a classification decision is processed at two levels. The first is to reject objects that are not contained in the model data base. The second is to identify the likelihood of error for classifications of known objects. Both stages are based on empirically determined thresholds of measures that are generated solely from the system's a priori knowledge, exploiting the known characteristics of both physical object space and feature space. Results are presented for a standardized object identification task, with a set of six similar known objects, and four unknown objects. It is shown that objects outside the model data base can be effectively rejected and that the accuracy of known object identifications can be increased by rejecting views of low classification quality. © 1989 IEEE