Model Structure Selection and Training Algorithms for an HMM Gesture Recognition System

Authors:
Nianjun Liu;Brian C. Lovell;Peter J. Kootsookos;Richard I. A. Davis
Affiliations:
University of Queensland;University of Queensland;University of Queensland;University of Queensland
Venue:
IWFHR '04 Proceedings of the Ninth International Workshop on Frontiers in Handwriting Recognition
Year:
2004

Citing 0
Cited 7

A belief-based sequential fusion approach for fusing manual signs and non-manual signals

Pattern Recognition
Handwritten character recognition using orientation quantization based on 3D accelerometer

Proceedings of the 5th Annual International Conference on Mobile and Ubiquitous Systems: Computing, Networking, and Services
A multi-class classification strategy for Fisher scores: Application to signer independent sign language recognition

Pattern Recognition
Hidden-Markov-model-based hand gesture recognition techniques used for a human-robot interaction system

HCII'11 Proceedings of the 14th international conference on Human-computer interaction: interaction techniques and environments - Volume Part II
A real-time bimanual 3D interaction method based on bare-hand tracking

MM '11 Proceedings of the 19th ACM international conference on Multimedia
Recognizing two handed gestures with generative, discriminative and ensemble methods via fisher kernels

MRCS'06 Proceedings of the 2006 international conference on Multimedia Content Representation, Classification and Security
Hidden Markov model for human to computer interaction: a study on human hand gesture recognition

Artificial Intelligence Review

Quantified Score

Hi-index	0.00

Visualization

Abstract

Hidden Markov models using the Fully-Connected, Left-Right and Left-Right Banded model structures are applied to the problem of alphabetical letter gesture recognition. We examine the effect of training techniques, in particular the Baum-Welch and Viterbi Path Counting techniques, on each of the model structures. We show that recognition rates improve when moving from a Fully-Connected model to a Left-Right model and a Left-Right Banded ýstaircaseý model with peak recognition rates of 84.8%, 92.31% and 97.31% respectively. The Left-Right Banded model in conjunction with the Viterbi Path Counting present the best performance. Direct calculation of model parameters from analysis of the physical system was also tested, yielding a peak recognition rate of 92%, but the simplicity and efficiency of this approach is of interest.