| Experimental
Computer Systems Laboratory
Computer Science Department State University of New York at Stony Brook |
ckyang@cs.sunysb.edu
www.cs.sunysb.edu/~ckyang Home: 631-928-7075 Lab : 631-632-8436 |
| Education | Ph.d. Candidate, Computer Science, (1995-Present)
State University of New York, Dissertation Topic: "On-the-Fly Processing of Compressed Volume Data" Advisor: Professor Tzi-cker Chiueh Master, Computer Science,
Bachelor, Mathematics,
|
| Research
Interests |
My research interests are mainly focused on the fields of Volume Visualization,
Computer Graphics and Multimedia Systems.
|
| Research
Summary |
My Ph.D. research here at Stony Brook involves mainly on volume visualization. And the goal is to address the two general problems in volume rendering: huge data sets and lengthy rendering process. The first one burdens the storage system as well as the run-time memory requirement while the second one hinders the interactivity thus making most rendering systems far from attractive or even practically useless. To address the first problem, compression is an obvious choice. However, the naive use of decompression then rendering only solves the storage issue while a more clever integration of both may help reducing the rendering time as well. This project attempted to solve both problems in one shot by making use of the FPST (Fourier Projection Slice Theorem), which can provide an asymptotically faster algorithm for volume rendering. By partitioning the data sets into sub-cubes and compositing the sub-images from applying the FPST on each of them, our system can perform rendering directly in the compression domain, thus saving data loading time, reducing memory footprint, and at the same time being asymptotically faster. For irregular grids, where the structure of data sets is quite complex and apparently the previous method cannot be applied, we therefore used a different strategy called on-the-fly rendering during decompression. The system, called Gatun, not only loads a data set from its compressed form, thus saving the data loading time, but also performs a garbage collection scheme to further minimize the working set of the renderer, thus further improving the rendering speed as a result. Both compression and rendering algorithms in Gatun exploit the same local connectivity information among adjacent tetrahedra, and thus can be tightly integrated into a unified implementation framework. In this project, we further incorporate volume simplification into Gatun. Since simplification is a form of lossy compression, the on-the-fly volume simplification algorithm provides a powerful mechanism to dynamically create versions of a tetrahedral mesh at multiple resolution levels directly from its losslessly compressed representation. With decompression, simplification and rendering all merged together, Gatun becomes a seamless and powerful pipeline for volume visualization. This work re-examines implementation strategies of the ray casting algorithm, taking into account both computation and I/O overheads. Specifically, we developed a data-driven execution model for ray casting that achieves the maximum overlap between rendering computation and disk I/O. It can also be readily extended to do out-of-core visualization as well. The main idea of our approach is to convert an application into two threads: a computation thread , which is the original program containing both computation and disk I/O, and a prefetch thread , which contains all the instructions in the original program that are related to disk I/O. At run time, the prefetch thread runs far ahead of the computation thread, so that blocks can be prefetched and put in the file system buffer cache before the computation thread needs them. A source-to-source translator is developed to automatically generate the prefetch and computation thread from a given application program, without any user intervention. My master thesis in National Taiwan University involved mainly three techniques: We applied this technique to model stalactites, whose shapes are usually much sharper and simpler. Using fractal theory to employ randomly and adaptively recursive subdivision at each level, a good approximation can be achieved. Stalagmites are more difficult to model because of their much higher complexity. Therefore instead of using fractal subdivision, we simulate the natural process of erosion and accumulation, the reasons how they have been formed, to obtain a better approximation. The global illumination method radiosity (by progressive refinement) is used to provide a more photorealistic lighting environment for the synthesized cavern with stalactites and stalagmites. I have also been involved in the following project where I have designed most of the graphical user interfaces. To have a more efficient display and better control of the video components required from the project, I have traced and modified the well known Berkeley Mpeg Player (for video display) and XV (for colormap tuning) thus gaining substantial amount of knowledge of the X windows under the Unix environment. In this project, I have also implemented the video annotation part, whose enabling technique is the underlying object tracking in video mechanism that combines the algorithms of active contour (snake) and motion estimation. This system, called Zodiac provides users a conceptually clean and semantically powerful branching history model of edit operations to organize the authoring process, and to navigate among the design alternatives. In addition, by analyzing the edit history, Zodiac is able to reliably detect a composed stream's shot and scene boundaries, which facilitate interactive video browsing. It also features a video object annotation capability that allows users to associate annotations to moving objects in a video sequence. The annotations themselves could be text, image, audio, or video. Zodiac is built on top of MMFS (MultiMedia File System), a file system specifically designed for interactive multimedia development environments, and implements an internal buffer manager that supports transparent lossless compression/decompression. |
| Experience |
Built a generic graphics rendering system, which incorporates 3D transformations, clippings, different lighting and shadings. Designed a modeling system which combined fractal theory for subdivision and a natural simulation of erosion and accumulation effect for a cavern with stalactites and stalagmites. Studied and implemented the radiosity rendering method. Performed system maintenance for hundreds of machines, including both software and hardware setup and troubleshooting. Designed a large-scale documentary and personnel system using Gupta's SQLwindows. The system included detailed interface design and full-blown database functionalities. Studied the generic wavelet theory and how it is used to perform multiresolution analysis for arbitrary surface mesh type. Also implemented a generic 3D volume compressor that can support different wavelet transforms. For courses of "C & Unix" and "Computer Graphics". Involved in multiple projects. See "Research Summary". |
| Journal
Publications |
Tzi-cker Chiueh, Tulika Mitra, Anindya Neogi and Chuan-kai Yang, ACM Multimedia Systems Journal, special issue on Multimedia Authoring and Presentation Techniques, vol.8 no.3 2000. |
| Referred
Conference Publications |
Chuan-kai Yang, Tulika Mitra and Tzi-cker Chiueh, to appear in the FREENIX track of the Usenix 2002 annual conference, Monterey, CA, June, 2002. Chuan-kai Yang and Tzi-cker Chiueh, VisSym '01, Joint Eurographics - IEEE TCVG Symposium on Visualization, Ascona, Switzerland, May, 2001. Chuan-kai Yang, Tulika Mitra and Tzi-cker Chiueh, IEEE Visualization '2000, Salt Lake City, Utah, October 2000. Tulika Mitra, Chuan-kai Yang and Tzi-cker Chiueh, IEEE Multimedia 2000, New York City, July 2000. Tzi-cker Chiueh, Tulika Mitra, Anindya Neogi and Chuan-kai Yang ACM Multimedia Conference, September 1998. Tzi-cker Chiueh, Chuan-kai Yang, Taosong He, Hanspeter Pfister and Arie Kaufman, IEEE Visualization '97, Phoenix, AZ, October 1997. Ming Ouhyoung and Chuan-kai Yang, Proc. of CAD/Graphics'93, New Advances in Computer Aided Design and Computer Graphics, Beijing, PRC, August 1993. Chuan-kai Yang and Ming Ouhyoung, Proc. of 1992 International Computer Symposium, Taiwan, 1992. |
| Technical
Reports |
Chuan-kai Yang, Research Proficiency Exam Report, September 2000. Chuan-kai Yang and Tzi-cker Chiueh, March, 2001 (submitted for publication). Chuan-kai Yang, Dissertation Proposal Report, October 2001. Chuan-kai Yang, Phd Dissertation, August 2002. |
| Languages | C/C++, Java, (Visual) Basic, Tcl/Tk, OpenGL, C Shell, Scheme, ML, SmallTalk, SQLwindows, Fortran, Pascal, 80x86 Assembly, |
| Platforms | Linux, FreeBSD, Solaris, SunOS, Irix, Dos/Windows |
| References | (In alphabetical order)
Prof. Tzi-cker Chiueh (advisor) Department of Computer Science State University of New York at Stony Brook Stony Brook, NY 11794-4400 (O)+1-631-632-8449 (F)+1-631-632-8334 chiueh@cs.sunysb.edu Prof. Kwan-liu Ma
Prof. Klaus Mueller
Prof. Ming Ouhyoung
|