New for: D1, D2, D3, D4, D5
In this talk, I will introduce the typical distortions used in geometry processing, and how to reduce and control these distortions in order to generate natural-looking results in a few applications, including a) high level image editing. By embedding an image into a mesh formed of polygons, we can transform the image into another by deforming the mesh. Equipped with proper definition of the distortion, different image editing goals can be achieved, such as content-aware image resizing, which allows to change the image size while keeping its overall appearance undistorted, and rephotography, which allows to change a photo as if it was shoot from different view positions or cameras. b) key-frame animation. As one of the most important applications in computer graphics, modern geometry processing tools can greatly reduce the artist efforts in creating visually appealing animation sequences.