Parallel Optical Flow Using Local Voting
James J. Little, HH Bülthoff, Tomaso Poggio · International Conference on Computer Vision · 2005
We describe a parallel algorithm for computing optical flow from short-range motion. Regularizing optical flow computation leads to a forruulation which minimizes matching error and, at the same time, maximises smoothness of the optical flow. We develop an approximation to the full regularization computation in which corresponding points are found by comparing local patches of the images. Selection aniong competing matches is performed using a winner-take-all scheme. The algorithm accommodates many different image transformations uniformly, with siniilar results, from brightness to edges. The optical flow computed froni different image transformations, such as edge detection and direct brightness computation, can be simply combined. The algorithm is easily implemented using local operations on a finegrained computer, and has been implemented on a Connection Machine. Experiments with natural images show that the scheme is effective and robust against noise. The algorithm leads to dense optical flow fields ; in addition, inforniation from matching facilitates segmentation.
Cited by 52 works. doi:10.1109/ccv.1988.590023
Authors
References
- A Computational Approach to Edge Detection
- Determining optical flow
- Determining Optical Flow
- The Interpretation of Visual Motion
- Velocity determination in scenes containing several moving objects
- Vision: A computational investigation into the human representation and processing of visual information
- Determining Three-Dimensional Motion and Structure from Optical Flow Generated by Several Moving Objects
- On optimal nonlinear associative recall
- Directional selectivity and its use in early visual processing
- Displacement vectors derived from second-order intensity variations in image sequences