Optical flow estimation on Connection-Machine 2

Abstract
The gradient-based methods for optical flow estimation are based on a constraint equation which is defined for each image pixel. The structure of the constraint equation make the problem ill-posed so in the past solutions have been proposed some solutions based on regularization. On the contrary, under the assumption that in the immediate neighborhood of a pixel the optical flow field is smooth, the constraint equations in that neighborhood should have a common solution, in this case the problem is not ill-posed. Following this reasoning, an algorithm for evaluating the optical flow, which is suitable for parallel implementations of selected algorithms from the literature, for optical flow estimation, are presented with the intention of comparing their complexity and performance with respect to the proposed approach on a Connection Machine-2.

This publication has 14 references indexed in Scilit: