NOVEL STEREO IMAGE AND VIDEO RETARGETING APPROACH

摘要

In this paper, a novel stereo image and video retargeting approach is proposed. The proposed approach first extracts four feature maps: disparity map, gradient map, saliency map, and motion difference map. After that, the proposed approach constructs an importance map which combines the four feature maps by the weighted sum. Finally, the proposed approach constructs the target image using the seam carving method based on the importance map. Moreover, the proposed approach is extended to the retargeting of video sequences. Considering the images with time continuity, the proposed approach constructs an accumulated map by calculating the average of the importance maps for a period of time. The experimental results show that the proposed approach performs well in terms of the resized quality.