Realistic inverse lighting from a single 2D image of a face, taken under unknown and complex lighting

摘要: In this paper, we address a difficult inverse rendering problem with many unknowns: single 2D input image of an unknown face in environment, taken under conditions. First, the geometry and texture are estimated from image, using 3D Morphable Model. second step, considering superposition principle for light, estimate light source intensities as optimized non-negative weights linear combination synthetic illumination cone that face. Each is lighted by one directional non-lambertian reflectance non-convex geometry. Modeling lighting separately model enhances modeling analysis, provides information about environment face, facilitates realistic novel pose lighting.