Monocular Stereo Vision of Image Feature-aware Interactive Generation

Abstract

This paper presents an image feature extraction based monocular stereo vision method. According to the key feature-aware region of the image, the famous interactive segmentation method GrabCut is introduced. Combing with the image correction and affine transformation and stereoscopic generation, the method can get the resemble 3D vision effect include with the shadow generating and object folding from the original cut apart region in 2D space. The algorithm of image feature-aware monocular stereo vision is high effective and efficiency by experiment with several different background complexity samples. It’s readable and interesting in the usage of virtual reality and augment reality. Especially in the field of digital culture publication and naked eye 3D display, has great application value.