Unsupervised parameter selection for gesture recognition with vector quantization and hidden markov models

Abstract

This article presents an investigation of a heuristic approach for unsupervised parameter selection for gesture recognition system based on Vector Quantization (VQ) and Hidden Markov Model (HMM). The two stage algorithm which uses histograms of distance measurements is proposed and tested on a database of natural gestures recorded with motion capture glove. Presented method allows unsupervised estimation of parameters of a recognition system, given example gesture recordings, with savings in computation time and improved performance in comparison to exhaustive parameter search. © 2011 IFIP International Federation for Information Processing.