Evaluation and Comparison of Optical Flow Techniques
In collaboration with John Barron and Steven Beauchemin (University of
Western Ontario), I conducted an extensive performance analysis of
well-known optical flow algorithms (circa 1993).
With independent implementations and a common set of synthetic
and natural inputs, we showed that the results of different methods
differed significantly. Phase-based and gradient-based techniques
produced the most reliable results over the input sequences we used.
Our test suite and results, along with our implementations, have been
available to others on the web since 1994, and they are used widely.
They remain the benchmark against which new methods are judged.
- Barron, J.L., Fleet, D.J., and Beauchemin, S. (1994) Performance of
optical flow techniques. International Journal of Computer Vision,
- Barron, J.L., Fleet, D.J., Beauchemin, S., and Burkitt, T. (1992)
Performance of optical flow techniques. IEEE Conference on
Computer Vision and Pattern Recognition, Champaign, June, pp. 236-242
Return to David Fleet's home page.