Method for processing multi-viewpoint depth video

Abstract

The invention discloses a method for processing a multi-viewpoint depth video. The method comprises the steps as follows: recombination conversion operation is executed aiming at the depth video; secondly, the converted depth video is smoothly processed in a time domain; thirdly, recombination conversion operation is executed aiming at the smoothly processed depth video to obtain a pretreated depth video; fourthly, coding compression, decoding recombination operation are executed aiming at the pretreated multi-viewpoint depth video; fifthly, depth recovery and null field smooth processing are executed aiming at decoded and recombined multi-viewpoint depth video; and lastly, a virtual viewpoint video image is mapped by the processed multi-viewpoint depth video. The method has the advantages that the pretreatment method can improve the time domain relativity of a depth video sequence, and further effectively improve the coding efficiency of the depth video; as for the situation of a code QP equal to 22, 27, 32 and 37, the coding efficiency can be saved by 17.07 percent to 38.29 percent, and the mapping quality of the obtained virtual viewpoint video image can be improved by about 0.05 dB.

Description

technical field [0001] The invention relates to a video signal processing technology, in particular to a multi-viewpoint depth video processing method. Background technique [0002] The depth video sequence corresponds to the color video sequence, which reflects the distance between the object and the camera. The actual distance between each pixel in the color image and the camera is quantified to 0~255, and then this quantization value is assigned to the color image. The corresponding position becomes the depth map. The larger the brightness value in the depth video sequence, the closer the pixel is to the camera; otherwise, the farther the pixel is from the camera. Depth video sequences are captured by depth cameras or obtained by depth estimation. Since depth cameras are relatively expensive and have problems of application range and accuracy, current depth video sequences are generally obtained by depth estimation. [0003] The inaccurate calculation of depth will lead...

AI Technical Summary

Problems solved by technology

In addition, new noise will be introduced when the depth image passes through the H.264 encoder, and due to the special nature of the depth image, these encoding distortions are often most obvious in the edge area of ​​the depth, and the edge area is just the sensitive area for virtual viewpoint drawing, so Depth restoration techniques are needed to restore the real depth boundaries to improve the subjective and objective quality of virtual viewpoint rendering images

Images