Spatio-temporal Composite-Features for Motion Analysis and Segmentation
Motion estimation by means of spatio-temporal energy filters –velocity tuned filters– is known to be robust to noise and aliasing and to allow an easy treatment of the aperture problem. In this paper we propose a motion representation based on the composition of spatio-temporal energy features, i.e., responses of a set of filters in phase quadrature tuned to different scales and orientations. Complex motion patterns are identified by unsupervised cluster analysis of energy features. The integration criterion reflects the degree of alignment of maxima of the features’s amplitude, which is related to phase congruence. The composite-feature representation has been applied to motion segmentation with a geodesic active model both for initialization and image potential definition. We will show that the resulting method is able to handle typical problems, such as partial and total occlusions, large inter-frame displacements, moving background and noise.
KeywordsMotion Pattern Active Model Image Potential Previous Frame Integration Criterion
Unable to display preview. Download preview PDF.
- 2.Boykov, Y., Huttenlocher, D.: Adaptive bayesian recognition in tracking rigid objects. In: IEEE CVPR, vol. 2, pp. 697–704 (2000)Google Scholar
- 8.Simoncelli, E., Adelson, E.: Computing optical flow distributions using spatio-temporal filters. Technical Report 165, MIT Media Lab. Vision and Modeling, Massachusetts (1991)Google Scholar
- 12.Chamorro-Martínez, J., Fdez-Valdivia, J., García, J., Martínez-Baena, J.: A frequency domain approach for the extraction of motion patterns. In: IEEE Int. Conf. on Acoustics, Speech and Signal Processing, vol. 3, pp. 165–168 (2003)Google Scholar
- 20.Kovesi, P.: Invariant Measures of Image Features from Phase Information. PhD thesis, The University or Western Australia (1996)Google Scholar