Global Motion Fourier Series Expansion for Video Indexing and Retrieval

Authors:
E. Bruno;Denis Pellerin
Affiliations:
-;-
Venue:
VISUAL '00 Proceedings of the 4th International Conference on Advances in Visual Information Systems
Year:
2000

Citing 0
Cited 2

Video retrieval using spatio-temporal descriptors

MULTIMEDIA '03 Proceedings of the eleventh ACM international conference on Multimedia
Pattern recognition filtering and bidimensional FFT-based detection of storms in meteorological radar images

Digital Signal Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper describes a new framework for global motion feature extraction and presents a video indexing and retrieval application. Optical flow between two frames is expanded, directly from the image derivatives, in a Fourier series. This technique provides a good global motion representation over a few Fourier components. These Fourier components are relevant to discriminate complex motions, such as human activities. Results of indexing and retrieval on a database of human activities sequences are presented.