Interactive relighting editing method based on RGB-D image

Abstract

The invention discloses an interactive relighting editing method based on an RGB-D image. The method comprises the following steps of: preprocessing an input image to obtain a shadow-free reflectivity image, a bright-dark image and a shadow image; segmenting an original RGB image by using a MaskRCNN algorithm to obtain a segmentation mask of an object in a scene, and then obtaining a corresponding local depth map, a shadow-free reflectivity map and a shadow-free bright-dark map; decomposing the shadow-free bright-dark image into a global illumination detail image, a local illumination detail image and a spherical harmonic illumination image; sampling and visualizing the global spherical harmonic illumination image and the local spherical harmonic illumination image, and obtaining an adjusted global spherical harmonic illumination image and an adjusted local spherical harmonic illumination image through interactive editing; and synthesizing the global and local shadow-free reflectivity images, the adjusted spherical harmonic illumination image and the shadow image to obtain global and local RGB relighting images subjected to ambient light editing, and outputting the global and local RGB relighting images. The indoor complex scene image enhancement under a low-light condition can be effectively, conveniently and intuitively realized.

Description

technical field [0001] The invention relates to the fields of image processing technology and computer graphics technology, in particular to an interactive heavy-light editing method based on RGB-D images. Background technique [0002] Lighting information is an important factor that constitutes the quality of a photo. The lighting of a photo, especially the ambient light distribution, directly affects the visual quality of the photo. Especially in indoor low-light scenes, the photos taken have complex ambient light, and the shooting effect is often difficult to present directly in front of people. Even in professional photographers, they often need to use additional artificial Light source to improve the light distribution of the environment. However, due to the complexity of real scenes, it is difficult for professional photographers to achieve better lighting effects through post-editing. In addition, they cannot achieve natural and real local lighting editing to enhance...

AI Technical Summary

Problems solved by technology

[0004] The embodiment of the present application provides an interactive re-illumination editing method based on RGB-D images, which effectively solves the difficulty in the prior art to perform global and local effective illumination of complex scene images under indoor low-light conditions. enhanced question

Images