Jitter-Robust Video Retargeting With Kalman Filter And Attention Saliency Fusion Network
Hyunwoo Nam, Dubok Park, Kangwon Jeon · 2020
In this paper, we propose a novel framework for video retargeting with Kalman filter and Attention Saliency Fusion Network (ASFN). First, we estimate the importance map which accounts for a content-aware saliency map. The aspect ratio of a video is then resized by adaptive scaling function. Finally, the Kalman filter is applied to the image sequence for alleviating jittering and wavy effect of moving objects. Proposed method uses two consecutive images which can be implemented on real-time application. Experimental results validate the proposed framework can achieve content-preserving results while alleviating the jitter effect.