Stereoscopic Video Retargeting Based on Camera Motion Classification

Abstract

The existing stereo video retargeting algorithms commonly use a same methodology to perform resizing without considering different videos with various features, leading to the low quality of reconstructed videos. To address this issue, we propose a stereo video retargeting method based on camera motion classification, which employs different retargeting strategies to rescale stereo videos. We also design an adaptive stereo video classification method which determines the types of camera motion according to the distribution of motion vectors extracted from the left view of stereo videos. Besides, we develop a motion saliency detection method to eliminate the jittering of moving objects during video resizing. Experimental results show that the qualities of retargeted videos produced by our method are significantly superior to those of existing methods.