Image three-dimensional representation method based on parallax hierarchical segmentation, and system

Abstract

The invention discloses an image three-dimensional representation method based on parallax hierarchical segmentation, and a system. Firstly, parallax information of 3D images are obtained by utilizing an existing parallax estimation algorithm; secondly, depth direction quantized intervals are determined according to stereoscopic visual acuity of the visual sense and complexities of image contents in the depth direction, and different depth plane segmentation methods are established; a parallax map is segmented relying on depth intervals according to the quantized intervals determined by various depth planes, and image segmentation templates corresponding to the various depth planes are formed; and, finally, an original image is segmented by utilizing the image segmentation templates of the various depth planes, image contents obtained through segmentation are located on the corresponding depth planes and are redrawn on various planes of the corresponding depth spaces, and a depth space distributed three-dimensional representation manner corresponding to the image is formed. The original image is segmented based on the parallax information of the 3D images, and inner depth information of image objects in various depth intervals in the 3D images is significantly and visually representated.

Description

technical field [0001] The present invention relates to the field of stereoscopic image representation, in particular to stereoscopically presenting depth level information of stereoscopic images, especially a method and system for stereoscopic representation of images based on parallax layered segmentation. Background technique [0002] In recent years, the 3D industry has developed vigorously. Whether it is movies or television, 3D resources have become more and more widely used in people's daily lives. Due to the complexity of 3D resources and the discomfort of some viewers after watching 3D, such as dizziness and swollen eyes, more and more attention has been paid to basic research in the 3D field. As the basic technology of 3D, although the display technology of 3D content has matured, the current expression method cannot reflect the complete information of the image intuitively, especially as one of the most important characteristics of 3D images-- ---Depth informatio...

AI Technical Summary

Problems solved by technology

However, this representation can only quantitatively know the depth value characteristics of each pixel in the image, but it cannot qualitatively and intuitively perceive the depth characteristics from the color image.

Images