Robust Optical Flow with Combined Lucas-Kanade/Horn-Schunck and Automatic Neighborhood Selection

Abstract

Differential optical flow methods are widely used within the computer vision community. They are classified as being either local, as in the Lucas-Kanade method, or global, such as in the Horn-Schunck technique. Local differential techniques are known to have robustness under noise, whilst global techniques are able to produce dense optical flow fields. We will show that the Horn-Schunck Technique, when combined with Lucas-Kanade, can yield the advantage of having both robust and dense optical flow fields. Selection of neighborhood size is an important tuning parameter for the combined Lucas-Kanade/Horn-Schunck technique. Choosing the optimal neighborhood is a difficult task and greatly effects the performance of optical flow results. We outline a method for the automatic selection of neighborhood size based on Stein's Unbiased Risk Estimator (SURE). Algorithms are derived for a combined Lucas-Kanade/Horn-Schunck technique with automatic neighborhood selection. The performance of SURE neighborhood selection for the combined optical flow technique is simulated via Matlab, providing an illustration of the performance that is attainable.

Dense Estimation of Optical Flow Field Within the Mpeg-2 Compressed Domain

“…more than one candidate motion fits the function equally well, this ambiguity is known as the aperture problem. 3,4 In the above optical flow estimation process, the initial velocity estimation and MPEG constraint is used, which will do benefit to overcome aperture problem. However, false motion vectors and noisy motion vectors still must be removed to make the final motion vector field more accurate.…”

“…Among them, Lucas-Kanade method and Horn-Schunck method are most prominent. 3 The Lucas-Kanade method is a twoframe differential method, which is based on local Taylor series approximations of the image signal. The Horn-Schunck method is a global method which introduces a global constraint of smoothness to solve the aperture problem.…”

A VLSI system architecture for optical flow computation

“…Optical flow algorithms in general, estimate the changes of motion between two frames of an image with the assumption of constant brightness among subsequent frames. Most of the optical flow algorithms, however, require substantial computation and storage resources for computing the velocity of each pixel [3], [4], [5], [6]. A detailed comparison of these optical flow algorithms in terms of performance and efficiency can be found in [7] and [8].…”

An integrated sign language recognition system

“…The Lucas-Kanade method assumes small changes in motion in the scene between consecutive frames [13]. This assumption makes it possible to introduce a search window of known size around each pixel in the image, the bounds of which are assumed to be an upper limit to the vector of motion of the pixel within the scene.…”

“…It is assumed that the intensity of the pixels in the image are static, with only their location varying. This is expressed mathematically as [13]:…”