文献类型：期刊文献

英文题名：A MULTI-CUES BASED APPROACH FOR VISUAL SALIENCY DETECTION

作者：Zhang, Qiaorong[1];Wang, Yingfeng[1]

第一作者：张巧荣

通讯作者：Zhang, QR[1]

机构：[1]Henan Univ Econ & Law, Coll Comp & Informat Engn, 180 Jinshui East Rd, Zhengzhou 450046, Peoples R China

第一机构：河南财经政法大学计算机与信息工程学院

通讯机构：[1]corresponding author), Henan Univ Econ & Law, Coll Comp & Informat Engn, 180 Jinshui East Rd, Zhengzhou 450046, Peoples R China.|[1048412]河南财经政法大学计算机与信息工程学院;[10484]河南财经政法大学;

年份：2021

卷号：17

期号：4

起止页码：1435-1446

外文期刊名：INTERNATIONAL JOURNAL OF INNOVATIVE COMPUTING INFORMATION AND CONTROL

收录：EI(收录号：20213710891617);Scopus(收录号：2-s2.0-85114697859);WOS:【ESCI(收录号:WOS:000661271400021)】；

基金：This work is partially supported by the Key Scientific Research Project of Henan Higher Education Institutions (No. 18B520005). The authors also gratefully acknowledge the helpful comments and suggestions of the reviewers, which have improved the presentation.

语种：英文

外文关键词：Visual saliency; Saliency map; Background prior; Object estimation

摘要：Visual saliency detection is very useful in image and video processing. A visual saliency detection approach based on multiple cues is proposed in this paper. Firstly, we preprocess and decompose the original image into super pixels. The feature difference between each super pixel and other super pixels is calculated to get the feature contrast saliency map. Secondly, we extract the background regions from the edge regions of the image and calculate a background based saliency map according to the background regions. Thirdly, we estimate the foreground objects based on the background based saliency map, and calculate the foreground based saliency map. Finally, we integrate the feature contrast saliency map, background based saliency map and foreground based saliency map to get the final saliency map. Evaluations on four benchmark datasets show that the proposed algorithm outperforms other state-of-the-art algorithms on these datasets.