A VARIATIONAL FRAMEWORK FOR SIMULTANEOUS MOTION AND DISPARITY ESTIMATION IN A SEQUENCE OF STEREO IMAGES

In this paper, we present a variational framework for joint disparity and motion estimation in a sequence of stereo images. The problem involves the estimation of four dense fields: two motion fields and two disparity fields. In order to reduce computational complexity and improve estimation accuracy, the two motion fields, for the left and fight sequences, and the disparity field of the current stereo pair are jointly estimated, using the stereo-motion consistency constraint. In the proposed variational framework, the joint estimation problem is formulated as a convex programming problem in which a convex objective function is minimized under specific convex constraints. This minimization is achieved using an efficient parallel block-iterative algorithm. Experimental results involving real stereo sequences indicate the feasibility and robustness of our approach.