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.
https://hal.inria.fr/hal-01597029 Contributor : Hal IfipConnect in order to contact the contributor Submitted on : Thursday, September 28, 2017 - 10:24:44 AM Last modification on : Tuesday, October 10, 2017 - 1:32:29 PM Long-term archiving on: : Friday, December 29, 2017 - 2:24:34 PM
Przemysław Głomb, Michał Romaszewski, Arkadiusz Sochan, Sebastian Opozda. Unsupervised Parameter Selection for Gesture Recognition with Vector Quantization and Hidden Markov Models. 13th International Conference on Human-Computer Interaction (INTERACT), Sep 2011, Lisbon, Portugal. pp.170-177, ⟨10.1007/978-3-642-23768-3_14⟩. ⟨hal-01597029⟩