A saliency detection model using aggregation degree of color and texture
作者：
Highlights：
• We propose a saliency model using aggregation degree of color and texture.
• Aggregation degree and divergence are applied to extract color saliency maps.
• The local phase of hypercomplex is used to produce the texture saliency map.
• The final saliency map is reconstructed by fusing credit feature maps which are selected by an effective mechanism.
• The experiment shows that our method outperforms other state-of-art models.
摘要
Highlights•We propose a saliency model using aggregation degree of color and texture.•Aggregation degree and divergence are applied to extract color saliency maps.•The local phase of hypercomplex is used to produce the texture saliency map.•The final saliency map is reconstructed by fusing credit feature maps which are selected by an effective mechanism.•The experiment shows that our method outperforms other state-of-art models.
论文关键词：Saliency detection,Aggregation degree,Hilbert transform,Hypercomplex,Credit feature selection
论文评审过程：Received 15 March 2014, Revised 25 September 2014, Accepted 3 October 2014, Available online 30 October 2014.
论文官网地址：https://doi.org/10.1016/j.image.2014.10.004
