Discriminant appearance weighting for action recognition

Tetsu Matsukawa, Takio Kurita

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Extending popular histogram representations of local motion patterns, we present a novel weighted integration method based on an assumption that a motion importance should be changed by its appearance to obtain better recognition accuracies. The proposed integration method of motion and appearance patterns can weight information involving "what is moving" by discriminant way. The discriminant weights can be learned efficiently and naturally using two-dimensional fisher discriminant analysis (or, fisher weight maps) of co-occurrence matrices. Original fisher weight maps lose shift invariance of histogram features, while the proposed method preserves it. Experimental results on KTH human action dataset and UT-interaction dataset revealed the effectiveness of the proposed integration compared to naive integration methods of independent motion and appearance features and also other state-of-the-art methods.

Original languageEnglish
Title of host publication1st Asian Conference on Pattern Recognition, ACPR 2011
Pages7-11
Number of pages5
DOIs
Publication statusPublished - 2011
Externally publishedYes
Event1st Asian Conference on Pattern Recognition, ACPR 2011 - Beijing, China
Duration: Nov 28 2011Nov 28 2011

Publication series

Name1st Asian Conference on Pattern Recognition, ACPR 2011

Other

Other1st Asian Conference on Pattern Recognition, ACPR 2011
CountryChina
CityBeijing
Period11/28/1111/28/11

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition

Fingerprint Dive into the research topics of 'Discriminant appearance weighting for action recognition'. Together they form a unique fingerprint.

Cite this