Texture Adaptation for Human Tracking Using Statistical Shape Model

We propose a texture model construction method for human tracking based on a statistical texture-free shape model (distance map). In practical vision systems, a priori knowledge of the target object is often limited. In human tracking, since human figures involve a variety of images due to different clothes, the detected pixel values (color, brightness) can vary significantly from person to person. Accordingly, we cannot assume textures for the human figures, which makes it difficult to directly utilize pixel values in human tracking. In this paper, we describe a method to acquire texture models by detecting human position based on a texturefree shape model. Using our method, objects having a variety of pixel values can be tracked properly with a simple appearance model. Experimental results show the effectiveness of the proposed approach.