Research on salient object detection algorithm

Salient object detection can locate and segment salient objects on an image, and then concentrate limited processing resources on fewer areas of the image. The ability of salient object detection algorithms to eliminate redundant information has made it a hot topic. In this paper, the study of salient object detection algorithms in twodimensional still images is reviewed, and its development history and the advantages and disadvantages of related literatures are reviewed. At present, most of the salient object detection algorithms with excellent performance depend on a great variety of neural networks. However, there is still room for improvement. After that, the commonly used performance evaluation methods and related data sets of the present salient object detection algorithms are given. Finally, the future research directions are given. © 2018 Indian Pulp and Paper Technical Association. All rights reserved.