Image Visual Saliency Detection Algorithm Based on Local and Global Features
Weigang An, Pan Jin-xiao · International Journal of Applied Mathematics & Statistics/International journal of applied mathematics and statistics · 2013
The paper has proposed a novel image visual saliency detection algorithm based on local and global characteristic features to detect natural images in CIELab colorful space. The algorithm has divided the image into sub-blocks of 8×8. It is characterized in calculating a plurality local and global scale characteristics of each sub-block. To determine the weighted combination of the salient degree of sub-blocks, we can obtain the salient feature of the whole image. By calculating the contrast of the color chrominance, the four channels on the object have to obtain the salient edges. The salient features of image edges and salient objects in the image are obtained by the consolidated salient goals. The experimental results in the natural images are shown that the proposed algorithm in the paper can extract the image salient goals quickly, clearly and accurately.