PROJECT TITLE :

Robust Video Object Co segmentation - 2015

ABSTRACT:

With ever-increasing volumes of video knowledge, automatic extraction of salient object regions became even more significant for visual analytic solutions. This surge has conjointly opened up opportunities for taking advantage of collective cues encapsulated in multiple videos during a cooperative manner. However, it conjointly brings up major challenges, such as handling of drastic appearance, motion pattern, and pose variations, of foreground objects with indiscriminate backgrounds. Here, we tend to present a co segmentation framework to find and phase out common object regions across multiple frames and multiple videos in a joint fashion. We have a tendency to incorporate 3 types of cues, i.e., intraframe saliency, interframe consistency, and across-video similarity into an energy optimization framework that does not build restrictive assumptions on foreground appearance and motion model, and does not require objects to be visible in all frames. We conjointly introduce a spatio-temporal scale-invariant feature remodel (SIFT) flow descriptor to integrate across-video correspondence from the traditional SIFT-flow into interframe motion flow from optical flow. This novel spatio-temporal SIFT flow generates reliable estimations of common foregrounds over the entire video data set. Experimental results show that our method outperforms the state-of-the-art on a brand new in depth knowledge set (ViCoSeg).