Gray scale potential: A new feature for sparse image

Abstract Image features are the properties of image which can be used in image registration, image segmentation, target recognition and other fields of image processing. It plays an important role in image processing. In this paper, we first propose the gray scale potential (GSP) of image according to the relative position among the pixels. Second, we highlight the definition of GSP and how to calculate the GSP by taking the example of the binary images. Then we point out that the GSP can reflect an intrinsic feature of image. As for binary image, it reflects the relative distances of pixels to a baseline or to a reference point, and if the image is gray image, it reflects not only the distances but also the gray level feature. The GSP has obvious advantage in representing the sparse image, because it can reduce the computational work and storage. Finally, some experimental results are given to illustrate the characters of GSP. It shows that the GSP of image is a steady feature and can be used for target recognition.