Visual saliency detection based on region descriptors and prior knowledge
作者：
Highlights：
• We proposed a visual saliency method based on region descriptors and prior knowledge of saliency distribution.
• The region descriptors are introduced to better characterize the attribute of image segments.
• The prior knowledge of saliency distribution is exploited to heighten the contrast between salient region and background.
• Our method highlights the salient region uniformly and suppresses the background noise.
• The experiment shows that our method outperforms other state-of-art models.
摘要
Highlights•We proposed a visual saliency method based on region descriptors and prior knowledge of saliency distribution.•The region descriptors are introduced to better characterize the attribute of image segments.•The prior knowledge of saliency distribution is exploited to heighten the contrast between salient region and background.•Our method highlights the salient region uniformly and suppresses the background noise.•The experiment shows that our method outperforms other state-of-art models.
论文关键词：Visual saliency,Visual attention,Object detection,Nonlinear feature integration
论文评审过程：Received 12 July 2013, Revised 7 December 2013, Accepted 16 January 2014, Available online 30 January 2014.
论文官网地址：https://doi.org/10.1016/j.image.2014.01.004
