Bootstrap optical flow confidence and uncertainty measure

Abstract

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues.Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. a b s t r a c tWe address the problem of estimating the uncertainty of optical flow algorithm results. Our method estimates the error magnitude at all points in the image. It can be used as a confidence measure. It is based on bootstrap resampling, which is a computational statistical inference technique based on repeating the optical flow calculation several times for different randomly chosen subsets of pixel contributions. As few as ten repetitions are enough to obtain useful estimates of geometrical and angular errors. For demonstration, we use the combined local-global optical flow method (CLG) which generalizes both LucasKanade and Horn-Schunck type methods. However, the bootstrap method is very general and can be applied to almost any optical flow algorithm that can be formulated as a pixel-based minimization problem. We show experimentally on synthetic as well as real video sequences with known ground truth that the bootstrap method performs better than all other confidence measures tested.

Enhancement of imagery of objects with highly dynamic brightness and large rotational motion

“…Naturally, all error estimations assume presence of some type of ground truth (image or OF field). Since that data is rarely available, various error measures have been developed [19][20][21][22][23][24]. Most of the developed measures addressed the two former error metrics.…”

Sparse Aggregation Framework for Optical Flow Estimation

“…where β m (x) is a confidence measure associated to w m (x). Apart from [14,15], existing confidence measures are dedicated to specific motion estimation methods. For a variational approach, [7] uses the inverse of the global energy.…”

Application of immune evolutionary algorithm to determine the position and rotation of parallel manipulator

“…As for these methods, optical flow based methods [1][2][3] are sensitive to errors in optical flow; Gradient-based methods [4][5] are generally based on the brightness change constraint equation using least-squares method as estimator, which is sensitive to the residual errors that are not Gaussian distributed. The feature-based approaches [6][7][8] are usually very accurate, but the success of these methods depends on the relative arrangement of the feature points and the camera.…”