zbMATH — the first resource for mathematics

A real-time region-based motion segmentation using adaptive thresholding and \(k\)-means clustering. (English) Zbl 1052.68738
Stumptner, Markus (ed.) et al., AI 2001: Advances in artificial intelligence. 14th Australian joint conference, Adelaide, Australia, December 10–14, 2001. Proceedings. Berlin: Springer (ISBN 3-540-42960-3). Lect. Notes Comput. Sci. 2256, 213-224 (2001).
Summary: This paper presents an approach for a real-time region-based motion segmentation and tracking using an adaptive thresholding and \(k\)-means clustering in a scene, with focus on a video monitoring system. In order to reduce the computational load to the motion segmentation, the presented approach is based on the variation regions application of a weighted \(k\)-means clustering algorithm, followed by a motion-based region merging procedure. To indicate motion mask regions in a scene, instead of determining the threshold value manually, we use an adaptive thresholding method to automatically choose the threshold value. To image segment, the weighted \(k\)-means clustering algorithm is applied only on the motion mask regions of the current frame. In this way we do not to process the whole image so that the computation time is reduced. The presented method is able to deal with occlusion problems. Results show the validity of the presented method.
For the entire collection see [Zbl 0984.00055].
68T45 Machine vision and scene understanding
68U10 Computing methodologies for image processing
Full Text: Link