More about HKUST
Exploiting the Richness of Digital Photographs and Videos
Speaker: Dr. Sylvain PARIS Computer Graphics Group Computer Science and Artificial Intelligence Lab. Massachusetts Institute of Technology Title: "Exploiting the Richness of Digital Photographs and Videos" Date: Friday, 25 May 2007 Time: 4:00pm - 5:00pm Venue: Room 4204 (CSE Graphic Lab, via. 19) HKUST Abstract: Cameras have become popular with the recent development of digital equipment. Today, it is extremely simple to take high-resolution pictures and movies, thereby giving easy access to an abundance of information. However, the size and number of acquired images challenge most existing algorithms. In this context, I will present my work on computational photography and low-level image processing. I will first expose a method to transfer the visual qualities of an artist's photograph to another picture. This method makes it easy for novice photographers to achieve compelling renditions and can also improve the workflow of professionals. In the second part, I will describe the bilateral filter, a nonlinear edge-preserving process which is becoming ubiquitous in computational photography. I will reformulate this filter in a higher-dimensional homogeneous space and show that, using signal-processing arguments, it is possible to very quickly compute an approximation visually similar to the exact result. Finally, I will briefly present my on-going projects that extend these ideas to image segmentation and video processing. *********************** Biography: Sylvain PARIS is currently working with Frmdo Durand as a post-doc in the Computer Graphics Group in the Computer Science and Artificial Intelligence Laboratory at the Massachusetts Institute of Technology. His work mainly focuses on extracting information from real images. The goal is to obtain data useful for Computer Graphics, i.e. suitable for rendering new images. During his PhD with Francois Sillion, he developed new solutions to face relighting, shape reconstruction from image sequences, and for recovery of the 3D geometry of hair. The main results of this work are an optimal complexity algorithm for acquiring precise 3D models from several photographs, and a technique that exploits a video sequence to build a dense set of 3D lines that match someone's hairstyle.