STEM-END/CALYX DETECTION IN APPLE FRUITS Comparison of Feature Selection Methods and Classifiers

Abstract

A multiple classifier system to localize stem-ends and calyxes of apple fruits was introduced previously. In this paper we not only introduce a new decision step to this system, but also provide comparisons of several feature selection algorithms and classifiers used. Our results prove that floating forward selection is the best within heuristic methods and support vector machines are better than nearest neighbor classifier in discriminating stem-ends/calyxes from defects.