Saliency ranker: A new salient object detection method

Zun Li, Congyan Lang, Songhe Feng, Tao Wang

Abstract: Highlights•A novel learning-to-rank model for salient object detection is proposed.•The model formulates saliency detection as a cost-sensitive label ranking problem.•The model ranks saliency values in a descending order to fit the ordinal correlation.•A low-rank matrix recovery theories is introduced to optimize the proposed model.•A trace norm regularization is used to control the relevance between saliency labels.