Performance evaluation of salient object detection techniques

Abstract Recently, the detection and segmentation of salient objects that attract attention human visual in images is determined by using object (SOD) techniques. As an essential computer vision problem, SOD has increasingly attracted researchers’ interest over years. While a lot models applications have been proposed, there still lack deep understanding issues achievements. A comprehensive study on recent techniques provided this paper. Precisely, paper presents review from various perspectives. Various image are presented such as based machine learning or learning, second perspective concentrates classifying them into supervised unsupervised last one manual approach, semi-automatic fully automatic approach so on. Then, summarization datasets used for SOD. Finally, analyses comparison results presented.