Dense optical flow from multiple sparse candidate flows using two pass dynamic programming

Optical flow forms an important initial processing stage for many machine vision tasks. A framework is presented for the recovery of dense optical flows from image sequences containing large motions. Sparse feature correspondences are used to assign multiple candidate optical flows to each image pixel. This set of flows is then augmented with additional perturbed flows to allow for non-rigid motions. An energy functional comprising of a matching term and smoothness term is then minimized using a two pass dynamic programming algorithm to produce a final smooth optical flow field. The proposed algorithm shows a clear increase in recovered optical flow accuracy when compared to a hierarchical approach and a brute force block matching approach of similar computational complexity.