A complete confidence framework for optical flow

Abstract

Assessing the performance of optical flow in the absence of ground truth is of prime importance for a correct interpretation and application. Thus, in recent years, the interest in developing confidence measures has increased. However, by its complexity, assessing the capability of such measures for detecting areas of poor performance of optical flow is still unsolved. We define a confidence measure in the context of numerical stability of the optical flow scheme and also a protocol for assessing its capability to discard areas of non-reliable flows. Results on the Middlebury database validate our framework and show that, unlike existing measures, our measure is not biased towards any particular image feature. © 2012 Springer-Verlag.