Bi-subspace saliency detection
Jin Gui Lu, Fei Dou · 2017
Relevant to visual attention mechanism, visual saliency are concerns of detecting the salient regions in image and video. Saliency detection is a critical pre-process in many computer vision applications, which requires high-level visual recognition and scene understanding as well as low-level photo and video processing. In this paper, We propose a new bi-subspace bottom-up saliency detection model. Inspired by visually intuitions and bi-subspace priors, we formulate the problem as subspace analysis depicting the background aside from the saliency object in the image. In the meanwhile, another subspace representation via a group-sparse constraint is proposed to depict the structure of the object in our model. Numerical experiment results show that our method is more stable and accurate than other methods.