More about HKUST
Image-Space Acceleration for Real-Time Rendering
PhD Qualifying Examination Title: "Image-Space Acceleration for Real-Time Rendering" Mr. Lei YANG Abstract: Recent advances of high-performance graphics hardware and relevant algorithms have made real-time rendering of many photorealistic special effects and complex scenes possible. As performance is one of the primal demands of such real-time rendering systems, many researchers have focused on general and special techniques for accelerating the process, resulting in a multitude of different algorithms. In particular, algorithms that are based on or manipulate images receive special attention, due to their simplicity in both representation and implementation. This report gives an overview of the present image-space techniques used for accelerating general real-time rendering tasks. The techniques are categorized into four basic acceleration methodologies: simplifying data, removing futile data, reusing duplicated computation, and trading quality for speed. For each of the types, relevant techniques from the groundbreaking ones to the state-of-the-art ones are reviewed in details, grouped by their problem-solving strategies. The strengths and weaknesses of the methods are also discussed, together with their relationship to other approaches. The effectiveness of the techniques in discussion varies dramatically when applied to different applications. Each technique has its own constraints on scene complexity, type of motion, hardware support, and bottleneck type. At the end of this report, we will review and compare different approaches in a table. Their targets, limitations and artifacts are highlighted for facilitating technique selection as well as exposing promising areas for algorithmic improvements. Date: Tuesday, 26 August 2008 Time: 4:00p.m.-6:00p.m. Venue: Room 3304 lifts 17-18 Committee Members: Dr. Pedro Sander (Supervisor) Dr. Chi-Keung Tang (Chairperson) Dr. Huamin Qu Prof. Long Quan **** ALL are Welcome ****