Real-shot double-viewpoint 3D video optimization method and system oriented to naked 3D display

Abstract

The invention discloses a real-shot double-viewpoint 3D video optimization method and system oriented to naked 3D display. The method comprises the following steps: inputting a real-shot double-viewpoint 3D video; selecting a left picture or a right picture as a first viewpoint image, and a corresponding right picture or left picture as a second viewpoint image; partitioning the first viewpoint image into basic computing units; searching for matching units in the second viewpoint image based on the basic computing units of the first viewpoint image to compute a parallax field; converting the parallax field based on the basic computing units into a parallax field based on pixels by an edge-directed interpolation method; and normalizing the parallax field based on the pixels, and outputting the first viewpoint image and the parallax field based on the pixels in an image and depth 3D video format. Through adoption of the real-shot double-viewpoint 3D video optimization method and system, complicated later processing can be avoided in order to eliminate differences between the left picture and the right picture of the real-shot double-viewpoint 3D video. Through a semi-automatic parallax computing method, depth information or parallax information of the real-shot double-viewpoint 3D video can be extracted effectively, and the 3D effect of the real-shot double-viewpoint 3D video is kept.

Description

technical field [0001] The invention belongs to the technical field of 3D video optimization, and relates to a 3D video optimization method, in particular to a real-shot dual-viewpoint 3D video optimization method for naked-eye 3D display; at the same time, the invention also relates to a real-shot for naked-eye 3D display Dual-view 3D video optimization system. Background technique [0002] Dual-view 3D video is a general 3D video format and the main form of 3D video content. Dual-view 3D video can be captured by dual-camera 3D cameras, which is the main way to obtain 3D video content. However, the dual-camera 3D camera uses dual cameras to capture the left and right images respectively, resulting in differences between the left and right images of the real-shot dual-view 3D video, such as inconsistent brightness, inconsistent color, and vertical parallax, which will significantly affect the 3D image of the real-shot dual-view 3D video. Therefore, real shooting of dual-vi...

AI Technical Summary

Problems solved by technology

[0004] Glasses-free 3D display devices that play 3D film and television and other well-post-processed dual-viewpoint 3D videos have appeared and are widely used. However, the naked-eye 3D display devices that directly play real-shot dual-viewpoint 3D videos face many problems: (1) real-shot There are problems such as brightness inconsistency, color inconsistency, and vertical parallax in the left and right images of dual-viewpoint 3D videos that significantly affect the 3D effect; Inaccurate parallax extraction, leading to serious problems such as 3D image tearing

Images