Unsupervised image segmentation using defocus map and superpixel grouping

Image segmentation is an important and difficult issue in computer vision and image processing. It is categorized into two categories, supervised image segmentation and unsupervised image segmentation. The supervised method are not convenient since it needs the interactions of users. In this paper, we proposed an unsupervised method. It uses a defocus map, edge and color as similarity attributes of pixels or superpixels to generate an edge strength map. Then, we construct a minimum spanning tree with the superpixels and the edge map to divide the image to the foreground and background. In our experiment, our method doesn't need user interaction and the performance is better than previous superpixels grouping methods.