Saliency detection based on multi-cue and multi-scale with cellular automata
Ling Huang, Songguang Tang, Jiani Hu, Weihong Deng · 2016
Saliency detection plays an important role in computer vision. This paper proposes a saliency detection algorithm which is based on multi-cue and multi-scale with cellular automata. The algorithm constructs a background-based map at first and optimizes it with an automatic updating mechanism - single-layer cellular automata. Furthermore, two important visual cues, focusness and objectness, are added to evaluate saliency in different perspectives. In addition, multi-scale is introduced to avoid the saliency results' sensitive to different scales and the output saliency map is generated by multi-layer fusion. Extensive experiments on three public datasets comparing with other state-of-the-art results demonstrate the superior of the algorithm.