Depth map generation device

Abstract

The invention provides a depth map generation device, including an input module used for setting video synchronization signals, user variable input parameters and multi-channel image data, a gray scale conversion module used for calculating gray scale data of each pixel; a minimum depth path acquisition module used for generating a minimum depth path according to the gray scale data, a depth refinement module used for acquiring refined depth data and smooth depth data according to the minimum depth path, a user variable setting module used for carrying out updating calculation on the smooth depth data according to user variable input parameters to generate final depth data; a synchronization signal processing module used for performing output synchronization processing on the video synchronization signals, and an output module used for setting output parameters, wherein the output parameters include final depth data of each pixel and the updated video synchronization signals. The device according to the embodiment of the invention has the advantages of fast calculation speed and ability of direct real time conversion.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a depth map generation device. Background technique [0002] Stereoscopic video has realistic and immersive visual effects, and has applications in many fields such as advertising, film and television, games, and sports displays, especially in the field of stereoscopic television, which has broad application prospects. During the production of stereoscopic video content, it is usually necessary to convert flat video to stereoscopic video. [0003] Usually, a depth map is firstly generated, and a stereoscopic video image of a desired viewing angle can be generated from the depth map. In the prior art, there are mainly three ways to generate a depth map, which are manual drawing, semi-automatic depth map generation, and fully automatic depth map generation. Among them, manual rendering mainly uses manual participation to segment images and then assign depths to different ...

AI Technical Summary

Problems solved by technology

Among them, manual rendering mainly uses manual participation to segment images and then assign depths to different regions. This method is highly dependent on manual participation, with high labor costs and huge time-consuming

Semi-automatic depth map generation is the use of human participation and computer assistance to complete the depth map drawing. At present, there are mainly algorithms based on machines and image segmentation. These methods have high computational complexity, poor real-time performance, and high manual assistance costs.

Fully automatic depth map generation is mainly calculated by computer, and the depth map is generated through image understanding. If real-time calculation is to be performed, the hardware configuration requirements are high, the device portability is poor, and the interface with household devices such as TVs or set-top boxes is complicated.

Images