A Research Of Inverse Lighting Problems

Inverse rendering, which recovers lighting, reflectance, and textures from photographs, is an emerging technique in computer graphics and computer vision in recent years. In computer graphics field, researchers use environment information to render realistic images. However, according to the methods used in computer graphics, inverse rendering techniques focus on getting environment information from images. Inverse lighting problem based on images is a branch of inverse rendering. Inverse lighting and re-lighting techniques can be used in many applications, for examples, virtual reality, augmented reality, image composition and face recognition.Rendering in computer graphics regards lighting, reflectance attribute of object surface, shape of the object and camera parameters as input data, while deems images rendered in the scene as unknown information to solve. Inverse rendering techniques take image as input data, combined with the known information of object shape and camera parameters, and recover environment lighting or reflectance attributes of the surface. How to present illumination and reflectance is one of the basic problems in both realistic rendering and inverse rendering. This thesis first gives a review and analysis of these basic problems.