Frame for demixing and encoding stereo video

Abstract

The present invention discloses a three-dimensional video layered coding frame. A left view point image is designed to be basic layer and a standard coding form is adopted. A right view point image is designed to be a reinforcement layer, and the right view point image is coded to be a plurality of reinforcement layers with a quality scalable technology. The video sequence of the basic layer is coded according to the standard coding method. The video sequence of the reinforcement layer is coded by the quality layered coding mode. The frame effectively uses the characteristic of the three-dimensional video coding and the advantage of the single view point quality layered coding network, sets a predicting mechanism for a key reference frame, and reduces the diffusion of the code stream truncation error. On the premise of keeping the three-dimensional video coding efficiency, a three-dimensional video stream, the quality of which can be arbitrarily scaled according to the actual network bandwidth, is provided, and the influence of the code stream truncation to the image quality of the three-dimensional video can be effectively prevented.

Description

technical field [0001] The invention belongs to the field of video coding and processing, and in particular relates to the research of quality layered coding algorithms in the stereoscopic video coding process. Background technique [0002] Stereoscopic video is a research hotspot and a new development direction in the current video field, and has broad application prospects in the future. However, since the stereoscopic video image implies the depth information of the scene, the direct cost of this advantage is a sharp increase in the amount of video data, making the storage and transmission of video data more difficult, so using space, time and Information redundancy between two viewpoints Efficient compression of stereoscopic video is the key to promote stereoscopic video applications. [0003] At present, the most commonly used stereoscopic video coding framework is the combination of motion compensation prediction and parallax compensation prediction, such as figure 1...

AI Technical Summary

Problems solved by technology

But the disadvantage is that it has no effective measures to prevent the prediction error caused by the truncated code stream of the enhancement layer, so the image quality of the enhancement layer will continue to decline after the code stream is truncated until a new I frame of the basic layer appears. E.g figure 2 Among them, E2-P(t+1) and E3-P(t+1) frames are truncated due to the code rate limitation, and the subsequent frames from

Images