Texture Separation for Saliency Detection of Image with Cluttered Background
Xiao Dong, Minghong Liao, Xing Hua Gao, Juncong Lin · 2015
To identify important objects in an image is a prerequisite in many applications. Thus saliency detection has been investigated a lot for a long time. However, existing methods tend to be misled when texture-like small patterns spread over the image. In this paper, we present a novel strategy to firstly remove these texture-like small patterns before saliency detection. The experimental results show that our strategy is quite effective in improvement of the saliency detection results.