An image segmentation algorithm based on neighborhood evidence field

Image segmentation plays a fundamental role in many computer vision applications and it is also a classical difficult problem in image processing. A neighborhood evidence field model is proposed for image segmentation based on the analyses on the drawbacks of the traditional image segmentation methods. In our proposed model and related approach, belief functions are used to model the pixels and their corresponding neighborhood. According to the evidence theory, the spatial information is used in the process of image segmentation. Our proposed segmentation algorithm can effectively suppress noise. Thus it can achieve better segmentation results for images, especially for the images with big noise.