PROJECT TITLE :

Spatiotemporal Saliency Detection for Video Sequences Based on Random Walk With Restart - 2015

ABSTRACT:

A completely unique saliency detection algorithm for video sequences based mostly on the random walk with restart (RWR) is proposed during this paper. We have a tendency to adopt RWR to detect spatially and temporally salient regions. Additional specifically, we tend to initial notice a temporal saliency distribution using the options of motion distinctiveness, temporal consistency, and abrupt amendment. Among them, the motion distinctiveness springs by comparing the motion profiles of image patches. Then, we use the temporal saliency distribution as a restarting distribution of the random walker. Also, we design the transition probability matrix for the walker using the spatial options of intensity, color, and compactness. Finally, we tend to estimate the spatiotemporal saliency distribution by finding the steady-state distribution of the walker. The proposed algorithm detects foreground salient objects faithfully, whereas suppressing cluttered backgrounds effectively, by incorporating the spatial transition matrix and the temporal restarting distribution systematically. Experimental results on various video sequences demonstrate that the proposed algorithm outperforms standard saliency detection algorithms qualitatively and quantitatively.