Speaker: Ping Tan
Title: Towards Automatic Visual Content Creation
Computer graphics has been very successful. However, an important problem still remains unsolved. High quality graphical content (such as 3D models and realistic images or animations) is difficult to create, which limits graphics to mass market products, such as games and movies. To make graphics a media for our daily communication, we must bridge the gap between ordinary people and graphics professionals, and make visual content easy to create for the general public. This talk introduces the speaker's recent work towards this goal.
The speaker takes a data-driven approach for media synthesis to allow a novice user to easily create high quality images simply by providing high level input, such as sketches and text labels. The algorithm automatically searches for suitable image elements, and generates realistic output images. The speaker takes similar approach to build a virtual fitting room system, where a virtual mirror (i.e. video screen) displays a user wearing a virtual garment. The animations of the virtual garment are synthesized by tracking the user’s skeleton and utilizing corresponding pre-recorded video clips of the actual garment worn by a model. The speaker also develops systems to automate the process of 3D content creation. The speaker builds an image-based modeling systems to model large scale outdoor scenes such as trees, buildings and streets and also develops photo-geometric algorithms to precisely capture both 3D shape and surface reflectance.