Salient areas detection using color volume
Guanghai Liu · 2016
salient regions detection is a very popular topic in pattern recognition. A novel method was presented to detect salient regions using color volume and edge in the manner of simulating visual attention mechanism. In order to detect salient regions, we have combined color volume and edge cues together and used them as the primary visual features. Color and edge are important cues that attract human's attention. It can enhance the broad outline of the saliency regions. The results of the experiment validate that the proposed method can efficiently detect salient regions and outperform some famous methods in term of precision/recall and F-measure metrics on MSRA1000 dataset and ECSSD dataset.