WestminsterResearch

Auto clustering for unsupervised learning of atomic gesture components using minimum description length

Walter, Michael and Psarrou, Alexandra and Gong, Shaogang (2001) Auto clustering for unsupervised learning of atomic gesture components using minimum description length. In: Proceedings of IEEE ICCV Workshop on Recognition, Analysis, and Tracking of Faces and Gestures in Real-Time Systems, 2001. IEEE Computer Society, USA, pp. 157-162. ISBN 0769510744

[img]
Preview
PDF
1209Kb

Official URL: http://dx.doi.org/10.1109/RATFG.2001.938925

Abstract

We present an approach to automatically segment and label a continuous observation sequence of hand gestures for a complete unsupervised model acquisition. The method is based on the assumption that gestures can be viewed as repetitive sequences of atomic components, similar to phonemes in speech, governed by a high level structure controlling the temporal sequence. We show that the generating process for the atomic components can be described in gesture space by a mixture of Gaussian, with each mixture component tied to one atomic behaviour. Mixture components are determined using a standard EM approach while the determination of the number of components is based on an information criteria, the Minimum Description Length.

Item Type:Book Section
Uncontrolled Keywords:Gesture Recognition, Automatic Segmentation, Automatic Labelling, Data Driven Model Acquisition, Model Order Selection, Minimum Description Length (MDL), Atomic Gesture Components, Unsupervised Learning
Research Community:University of Westminster > Electronics and Computer Science, School of
ID Code:1491
Deposited On:08 May 2006
Last Modified:11 Aug 2010 15:30

Repository Staff Only: item control page