Salient regions detection based on color features
Yanbang Zhang, Lei Guo, Gong Cheng · 2016
In the paper, we provide a model to detect salient regions by using color features. First, the original image is segmented into superpixels to reduce computational complexity and suppress the disturbance of noise. Second, we select the superpixels in the corner of the image as background prior, and then compute the color contrast features in both Lab color space and the opponency color space. Furthermore, the location information is considered. Next, we employ two-dimension entropy to evaluate the performance of salient maps, and choose appropriate features to fuse. Finally, experimental results are given to show that the proposed model outperforms the some existing models on salient region detection.