Automatic image segmentation based on saliency maps and fuzzy SVM
Qian Zhao, Yueli Hu, Jialin Cao · 2009
Image segmentation is an essential step in image analysis. In this paper, we propose a new technique for unsupervised segmentation of viewer's attention objects from natural images by using visual attention (saliency maps) and Fuzzy SVM(FSVM). Firstly, a rectangular window meaning region of attention (ROA) is created based on the saliency map and corner points from an image in order to learn the object and background colour and texture information, then FSVM precisely segment single object in the rectangular ROA. Furthermore, mathematical morphology is used to refine the extracted results. Experimental results demonstrate the effectiveness of the proposed approach.