Human action recognition using non-separable oriented 3D dual-tree complex wavelets

Authors:
Rashid Minhas;Aryaz Baradarani;Sepideh Seifzadeh;Q. M. Jonathan Wu
Affiliations:
Department of Electrical and Computer Engineering, University of Windsor, Ontario, Canada;Department of Electrical and Computer Engineering, University of Windsor, Ontario, Canada;School of Computer Science, University of Windsor, Ontario, Canada;Department of Electrical and Computer Engineering, University of Windsor, Ontario, Canada
Venue:
ACCV'09 Proceedings of the 9th Asian conference on Computer Vision - Volume Part III
Year:
2009

Citing 8
Cited 0

Coupled hidden Markov models for complex action recognition

CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
Two-Dimensional PCA: A New Approach to Appearance-Based Face Representation and Recognition

IEEE Transactions on Pattern Analysis and Machine Intelligence
Actions Sketch: A Novel Action Representation

CVPR '05 Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Volume 1 - Volume 01
TemporalBoost for Event Recognition

ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1 - Volume 01
Successive Convex Matching for Action Detection

CVPR '06 Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Volume 2
Actions as Space-Time Shapes

IEEE Transactions on Pattern Analysis and Machine Intelligence
Recovering human body configurations: combining segmentation and recognition

CVPR'04 Proceedings of the 2004 IEEE computer society conference on Computer vision and pattern recognition
Sampled-Data Design of FIR Dual Filter Banks for Dual-Tree Complex Wavelet Transforms via LMI Optimization

IEEE Transactions on Signal Processing - Part II

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper introduces an efficient technique for simultaneous processing of video frames to extract spatio-temporal features for fine activity detection and localization Such features, obtained through motion-selectivity attribute of 3D dual-tree complex wavelet transform (3D-DTCWT), are used to train a classifier for categorization of an incoming video The proposed learning model offers three core advantages: 1) significantly faster training stage than traditional supervised approaches, 2) volumetric processing of video data due to the use of 3D transform, 3) rich representation of human actions in view of directionality and shift-invariance of DTCWT No assumptions of scene background, location, objects of interest, or point of view information are made for activity learning whereas bidirectional 2D-PCA is employed to preserve structure and correlation amongst neighborhood pixels of a video frame Experimental results compare favorably to recently published results in literature.