Adaptive color- and edge-based image segmentation using fuzzy techniques

Abstract

In this contribution a new segmentation technique is presented that is able to segment moving objects in front of a static, but arbitrary background. The system acquires information about the background automatically and is also able to adapt to slight changes within the background, for example caused of changing illumination conditions, automatically. Our technique uses a combination of color- and edge-based segmentation and stores the respective fuzzy information pixelwise. The available knowledge is employed to yield a stable and reliable result using fuzzy techniques. © Springer-Verlag 2001.