Last updated : Sept. 19th, 97

1. Collision Detection (CRCG and URC)

Collision detection is a fundamental issue in robotics and virtual environments simulating the real physical world. Making use of separating vectors between objects and exploiting time and geometric coherences, we have developed an efficient algorithm to accurately detect collision between a large number of convex polyhedral objects in 3D space. The next goal is to detect collision between objects bounded with free-form surfaces.

2. Complex Virtual Environments (CRCG)

We have modeled a complex indoor virtual environment based on the two floors of CYC building housing the Department of Computer Science and Information Systems at HKU. This model consists of over 100,000 textured polygons. A VR walkthrough system incorporating a new invisible polygonal culling algorithm has been developed to allow the user to explore this virtual environment with real-time frame rate. Currently, the system runs on SGI Maximum IMPACT workstation. We plan to extend the research to complex outdoor environments as well as virtual environments consisting free-form surfaces. The research in this direction is being conducted in two on-going projects: Imaged-Based VR System and Virtual Brain. The latter is a joint effort with the Department of Pathology at HKU.

3. NURBS Sweep Surface Modeling (RGC)

Sweep surfaces cover a wide and important class of entities in geometric design. We aim at modeling sweep surfaces that can be represented as NURBS surfaces, and we are also interested in the robust computation of the moving frame used for modeling sweep surfaces. Other basic issues of NURBS surface modeling are also under investigation in this project.

4. Quadric Surface Intersection (RGC)

Intersection of quadric surfaces is an important operation in solid modeling systems. We have proposed a simple and robust algebraic approach to computing the intersection curve of two quadric surfaces, based on bi-rational transformations between an intersection curve of two quadrics and a cubic plane curve. The algorithm produces parameterization as well as topological classification of the intersection curve.

5. Scientific Visualization (URC)

In SV graphics techniques are used to assist users in visualizing complex data involved in scientific investigations, engineering applications, or medical imaging. We study issues in visualization of 3D medical volume data as well as 3D flow data arising from CFD. These include the classification of volume data and efficient volume rendering with texture mapping hardware. Currently we are interested in a unified approach to rendering surface models together with volume data. This calls for efficient conversion of surface models into volume representations to explore the capacity of the emerging real-time volume rendering graphics hardware. In a joint project with the Department of Orthopaedic Surgery at HKU, we are developing a system for interactive 3D visualization of human bone density distribution to assist medical doctors in surgery planning. In another joint project with the Department of Civil Engineering at HKU, we have developed a prototype system for flow data visualization, called VISJET. VISJET uses a variety of 3D graphics and animation techniques to display the output data of a powerful CFD program, called JETLAG, developed at the Department of Civil Engineering, for simulating sewage discharge in the waters of Hong Kong.