Tracking by Hierarchical Representation of Target Structure
Tracking of spatially extended targets with variable shape, pose and appearance is a highly challenging task. In this work we propose a novel tracking approach using an incrementally generated part-based description to obtain a specific representation of target structure. The hierarchical part-based representation is learned in a generative manner from a large set of simple local features. The spatial and temporal density of observed part combinations is estimated by performing statistics over temporally aggregated data. Detected stable combinations consisting of multiple simpler parts encompass local, specific structures, which can efficiently guide a spatio-temporal association step of coherently moving image regions, which are part of the same target. The concept of our approach is proved and evaluated in several experiments.
Keywordshierarchical representation edge segments spatial statistics temporal statistics tracking
- 1.Adam, A., Rivlin, E., Shimshoni, I.: Robust fragments-based tracking using the integral histogram. In: CVPR, vol. 1, pp. 798–805. IEEE Computer Society, Los Alamitos (2006)Google Scholar
- 2.Ramanan, D., Forsyth, D.A.: Finding and tracking people from the bottom up. In: CVPR, vol. 2, pp. 467–474. IEEE Computer Society, Los Alamitos (2003)Google Scholar
- 4.Fidler, S., Leonardis, A.: Towards scalable representations of object categories: Learning a hierarchy of parts. In: CVPR, pp. 1–8. IEEE Computer Society, Los Alamitos (2007)Google Scholar
- 5.Gao, Q.: Perceptual tracking of edge features. In: ICIP, vol. 1, pp. 958–962 (1994)Google Scholar
- 9.Bouchard, G., Triggs, B.: Hierarchical part-based visual object categorization. CVPR 1, 710–715 (2005)Google Scholar