Development of Saliency Map Algorithms for Enhanced Detection in Multiple Images of Complex Scenarios

Abstract

Only a group of pixels of an image belong to the salient object in an image or a video. Visual saliency detection models are used to detect important and useful pixels which contribute to the identification of a salient object in an image or a video. Salient objects in complex backgrounds and the ones that exhibit pose variations are hard to detect. The task of salient object detection involves the identification of the pixels that newlinebelong to the visually salient object in a single image. Similarly, the task of co-salient object detection is the detection of the pixels that belong to the common salient object newlinein multiple images. The objective of this research is to investigate the various dimensions of salient and co-salient object detection. newline