Three-dimensional video cropping method

Abstract

The invention discloses a three-dimensional video cropping method. The method includes: marking left viewpoint video image and a right viewpoint video image of a three-dimensional video as key framesor non-key frames; for the key frame, calculating the three-dimensional saliency map of the left viewpoint video image to obtain the coordinate position of the upper left corner of the optimal cuttingframe of the left viewpoint video image, and further obtaining the coordinate position of the upper left corner of the optimal cutting frame of the right viewpoint video image; however, for non-key frames, according to the coordinate position and the motion vector of the upper left corner of the optimal cutting frame of the previous frame, directly obtaining the coordinate position of the upper left corner of the optimal cutting frame of the current stereoscopic video image; and finally, according to the coordinate position, the width and the height of the upper left corner of the optimal cutting frame, obtaining a relocation stereoscopic video. The obtained relocation stereoscopic video can well reserve important significant semantic information, can well keep time domain consistency, and can effectively adjust the size of the stereoscopic video.

Description

technical field [0001] The invention relates to a video signal processing method, in particular to a stereoscopic video clipping method. Background technique [0002] With the rapid development of stereoscopic display technology, various stereoscopic display terminals with different stereoscopic display functions have also appeared widely. However, due to the wide variety of stereoscopic display When a certain stereoscopic video is displayed on different stereoscopic display terminals, the size of the stereoscopic video must be adjusted first to achieve a stereoscopic display effect. The traditional method of video scaling is by cropping or scaling by a fixed ratio, but doing so may reduce the content in the video or cause significant object deformation. [0003] For stereoscopic video, if the same video relocation method is directly used for the left-viewpoint video and the right-viewpoint video, that is, the left-viewpoint video and the right-viewpoint video in the stereo...

AI Technical Summary

Problems solved by technology

The traditional method of video scaling is by cropping or scaling by a fixed ratio, but doing so may reduce the content of the video or cause significant object deformation

[0003] For stereoscopic video, if the same video relocation method is directly used for the left-viewpoint video and the right-viewpoint video, that is, the left-viewpoint video and the right-viewpoint video in the stereoscopic video are processed separately, then the traditional seamcarving method needs left and right The two slits are consistent to avoid destroying the stereoscopic nature of the original video. In the time dimension, the slits between adjacent frames need to be consistent so that the redirected results will not produce jitter distortion; the grid deformation method requires Keep the mesh in the visually significant area as little as possible, and distribute the deformation to the non-visually significant area. The calculation of the grid deformation is relatively large, and the inconsistent deformation of the left and right grids will easily cause the loss of three-dimensional characteristics.

Images