Structure-adaptive SOM to classify 3-dimensional point light actors' gender

Abstract

Classifying the patterns of moving point lights attached on actor's bodies with self-organizing map often fails to get successful results with its original unsupervised learning algorithm. This paper exploits a structure-adaptive self-organizing map (SASOM) which adaptively updates the weights, structure and size of the map, resulting in remarkable improvement of pattern classification performance. We have compared the results with those of conventional pattern classifiers and human subjects. SASOM turns out to be the best classifier producing 97.1% of recognition rate on the 312 test data from 26 subjects.