Interactive Segmentation of Multiple Images

Abstract— In this paper, we propose an optimization model for interactive segmentation of multiple images. The user marks some sample pixels or objects in one or more images. Then, the method employs the samples as strong priors to automatically segment other input images. A good feature of the method is that the segmentation is highly controllable by the user, so that the user can easily obtain the kind of objects that he/she wants. The approach is especially effective for segmentation of a large collection of images that share similar features, where the user inputs some samples once and for all. We demonstrate the usefulness of the model to the segmentation of various collections of bioimages.