Stereoscopic image display system and depth map generation method thereof

Abstract

An stereoscopic image display system and a depth map generation method thereof are disclosed. The method includes the following steps. An input 2D image is obtained. A first mask is generated by marking a plurality of specific pixels corresponding to a specific color in the input two-dimensional image. A deep learning model is used to perform semantic segmentation processing on the input two-dimensional image to obtain a second mask. A final mask is generated according to the first mask and the second mask. A monocular depth estimation is performed on the input two-dimensional image to obtain an initial depth map of the input two-dimensional image. A target depth map of the input 2D image is obtained by adjusting the initial depth map using the final mask map.

Images