Authors:
            
                    Jung-Hsuan Wu
                    
                        
                                1
                            
                    
                     and
                
                    Suguru Saito
                    
                        
                                2
                            
                    
                    
                
        
        
            Affiliations:
            
                    
                        
                                1
                            
                    
                    Tokyo Institute of Technology, Japan
                
                    ; 
                
                    
                        
                                2
                            
                    
                    Tokyo Institute of Technology and Ochanomizu University, Japan
                
        
        
        
        
        
             Keyword(s):
            Image-based Modeling, Image-based Rendering.
        
        
            
                Related
                    Ontology
                    Subjects/Areas/Topics:
                
                        Applications
                    ; 
                        Computer Vision, Visualization and Computer Graphics
                    ; 
                        Geometry and Modeling
                    ; 
                        Image-Based Modeling
                    ; 
                        Image-Based Rendering
                    ; 
                        Interactive Environments
                    ; 
                        Pattern Recognition
                    ; 
                        Rendering
                    ; 
                        Scene and Object Modeling
                    ; 
                        Software Engineering
                    ; 
                        Virtual Tours
                    
            
        
        
            
                Abstract: 
                Shadows play an important role in the human perception of 3D geometry in photographs, and properly synthesized shadows produce a realistic rendering result. Shadows, especially hard shadows, define powerful constraints between a light source, occluders, and the surfaces onto which shadows are cast. We aim to create a pseudo 3D scene in which the light produces shadows that are the same as those in the input image by taking the shadow constraints into account. Our method requires only a few user annotations and an input image, with the objects and their cast-shadows segmented in advance. The segmentation and the user annotations are used to create a shadow-free texture and a 3D scene, and users can then interactively edit the image by moving the objects and changing the lighting. Our system produces realistic shadows in real-time while enabling interactive editing.