High-performance tracking system
In this paper, we describe how reliable SSD feature selection, feature tracking and feature monitoring can be realized and interleaved into a high-performance system with no special-purpose hardware. We consider image brightness and contrast changes in the tracking system which haven't been treated before. We find the decoupled system outperforms the usual coupled system. We perform this calculation at multiple levels of resolution, leading to an adaptive algorithm for tracking both slow and fast motions. A new interpretation of feature selection is based on the trade off between noise resistance and linearization error. The overcorrectness problem in feature monitoring is addressed.
KeywordsFeature Selection Image Noise Image Region Visual Tracking Minimum Eigenvalue
Unable to display preview. Download preview PDF.
- 1.A. Blake, R. Curwen, and A. Zisserman. Affine-invariant contour tracking with automatic control of spatiotemporal scale. In Proc. Internal Conf. on Computer Vision, pages 421–430. IEEE Computer Society Press, 1993.Google Scholar
- 2.J. Shi and C. Tomasi. Good features to track. In Proc. IEEE Conf. Comp. Vision and Patt. Recog., pages 593–600. IEEE Computer Society Press, 1994.Google Scholar
- 3.C. Tomasi and T. Kanade. Shape and motion from image streams: a factorization method, full report on the orthographic case. CMU-CS 92-104, CMU, 1992.Google Scholar
- 4.P. Anandan. A computational framework and an algorithm for the measurement of structure from motion. Int. Journal of Computer Vision, 2:283–310, 1989.Google Scholar
- 5.B. D. Lucas and T. Kanade. An iterative image registration technique with an application to stereo vision. In Proc. Int. Joint Conf. Artificial Intelligence, pages 674–679. 1981.Google Scholar
- 6.N. Papanikolopoulos, P. Khosla, and T. Kanade. Visual tracking of a moving target by a camera mounted on a robot: A combination of control and vision. IEEE Trans. on Robotics and Automation, 9(1), 1993.sGoogle Scholar