A Matching and Tracking Strategy for Independently Moving Objects

Abstract

We present a robust and inherently parallel strategy for tracking "corner" features on independently moving (and possibly non-rigid) objects. The system operates over long, monocular image sequences and comprises two main parts. A matcher performs two-frame correspondence based on spatial proximity and similarity in local image structure, while a (racier maintains an image trajectory (and predictor) for every feature. The use of low-level features ensures an opportunistic and widely applicable algorithm. Moreover, the system copes with noisy data, predictor failure, and occlusion and disocclusion of scene structure. Motion and scene analysis modules can then be built onto this framework. The algorithm is aimed at applications with small inter-frame motion, such as videoconferencing.

A new approach to corner matching from image sequence using fuzzy similarity index

“…The method was able to handle points and lines in a unified manner by relying on the same principles for determining their correspondence. Shapiro et al (1992) developed a corner tracker by tracing corners using the correlation of small patches around corners to track them across frames. This, however, had the drawback that many of the pixels in a patch around a corner lie in the background and as these pixels change from frame to frame, the method yielded incorrect results.…”

Active tracking of foveated feature clusters using affine structure

“…Correspondence of corners between frames is achieved using a variation of the algorithms proposed by Shapiro, Wang and Brady Shapiro et al 1992). Corners are tracked from frame to flame using a constant image-velocity Kalman filter (Bar Shalom and FortInann 1988): state vector:…”

Assessing the performance of corner detectors for point feature tracking applications

“…Regions (or 'blobs') [32,26,27] normally correspond to smooth surface patches. Tracking such regions is not always easy, since minor differences between frames (due to image noise or image motion) can lead to very different segmentation in consecutive image frames [26].…”