Here are the projects offered this academic year. Students are highly recommended to contact me after the project briefing to know more about the project before finalizing your choices.
If you want to propose your own project that you think I can supervise, please email me to arrange for a discussion about it.
CE project briefing will be held on 2002-06-06[4], and CSIS project briefing will be held on 2002-06-17[1]. CSIS students can visit http://www.csis.hku.hk/~kpchan/project2002/ for information.
If you are interested in past or other projects, here they are:
| Title | nStudents | CSIS | CogSc | CE |
|---|---|---|---|---|
| Attention! | 2 | * | * | |
| Where's the eye? | 2 | * | * | |
| Tap it! | 2 | * | * | |
| Gamelan score typesetting system | 2 | * | * | * |
This project is for Cognitive Science students (preferred) and CSIS students. The number of students taking this project is limited to two.
This is an open-ended project. Students are expected to do their own research on the problem and discuss with the supervisor on possible computer implementations that demonstrate the solutions.Although we can close our eyes, we cannot close our ears. Yet, human beings have the ability of controlling their focus on particular audio sources. When we are in a concert, we can focus our attention on how the violin plays in an orchestra. When we are in a noisy restaurant, we can focus on our friend's voice even if more than one person is talking at the same time.
What's amazing is not only our ability to focus on a particular audio source. What's more, we can control what to focus on without much difficulty. For example, suppose we are one of the members in an annual dinner in a restaurant. At the same table, Vivian, Carmen and Brenda may be talking about music, Ben, Shing and Eric about world cup, Joe and Iris about the courses they've taken, and Sara and Felix about places to dine. You, being the odd one, can often join the discussion on any of the topics by focusing your attention to the particular discussion. The question is, can a computer do the same as you can? If so, how? If not, why not?
Since there is no close-form answer to this question, students taking this project is expected to do their own research on the problem. Literature survey will comprise the first part of the project to identify the domain in which computers can help in the problem. Students are then expected to design and investigate into computers algorithms that can "attend to" a particular audio source. Appropriate experiments, ranging from statistical analysis to Turing test, should be devised to test the algorithms designed.
This project is for Computer Engineering students (preferred) and CSIS students. The number of students taking this project is limited to two.
This is a practical project with potential application for the Hong Kong Observatory. Students strong in engineering mathematics are perferred. (What are gradient, divergence, and curl?)
On playing back radar or satellite pictures in time sequence (e.g., that for Typhoon York), tropical cyclones appear as rotating masses of cloud or rain. The eye is not always there. Individual pictures could be very messy with no obvious sign of where the centre is. However, in time sequences, we can see the rotation and intuitively senses the centre of rotation. This is sometimes used to determine the position of the centre of tropical cyclones. While it sounds obvious, no ready-made computer algorithm exists to do it objectively. Also, it is not as easy as it sounds because in weak tropical cyclones, the centre is not always completely surrounded by cloud or rain and one has to figure out a way to infer from an incomplete picture.
The problem is: given a sequence of radar pictures sampled every six minutes, and a sequence of satellite pictures sampled every hour, find a efficient algorithm to find the center of the eye of a tropical cyclone if there is one.
Students taking the project will be briefed with the problem by representatives from the Hong Kong Observatory. They are also expected to acquire the basic meterological knowledge required to complete the projects independently. The first part of the project would be a literature survey on approaches to solve the problem. Design and implementation of algorithms will follow, and testing will be done using past typhoon images, with expert advice from the Observatory.
This is the latest animation sequence of satellite images. Is there any (potential) tropical cyclone here?
This project is for CSIS and Computer Engineering students. The number of students taking this project is limited to two.
This is mainly an implementation project.
When we hear a piece of music, we can usually infer its beat and tap our feets in a few beats' time. If a computer can induce the beat of a piece of music, it can use it to do synchronization with other systems such as a video player.
Given a waveform or spectral representation of a piece of music, (e.g., AIFF file, WAV file, MP3 file) the system will induce the beat of the music in real time so that musical information can be communicated with other systems. The induced beat can be indicated by displaying it as a tempo curve, or by using sound and light.
Students taking this project are expected to be fluent in programming. The standard problem-solving procedure will be followed: problem analysis, literature survey, design, implementation, testing.
This is a related problem:
What are the next numbers of the following sequences if the numbers indicate when a note starts sounding?
| Sequence | Content |
| A | 10, 20, 30, 40, 50, 60, 70, 80, ? |
| B | 10, 21, 30, 38, 49, 60, 71, 80, ? |
| C | 10, 21, 30, 38, 49, 60, 65, 71, 80, ? |
| D | 10, 21, 25, 30, 34, 38, 49, 60, 65, 71, 76, 80, ? |
What if you're given the numbers in a sequence one by one?
This project is for CSIS, Cognitive Science, and Computer Engineering students. The number of students taking this project is limited to two.
This is mainly an implementation project with support from the Department of Music. Students with some background in music are preferred, and are expected to have learnt, or will learn the basics of Balinese gamelan music.
Those who have enrolled, or will enrol, in the 6-credit course MUSI2004 University Gamelan 1 or is willing to attend the community gamelan (held every Monday night) are highly preferred.
The Indonesian gamelan is a set of gong-chime type of percussion instruments. The department of music has devised a way to transcribe Balinese gamelan music into a "box" (table-like) notation to help nonmusicians to learn the music. Currently, the transcriptions are typeset using LaTeX+MusiXTeX typesetting system. Yet, it is expected that sheet music using Western music notation, and possibly other forms, are needed in the future. A system that can run on Windows and Macintosh platforms is needed to enter, store, and typeset those music.
How does Balinese gamelan look like? Here are some of the instruments:
A photo (courtesy Dorothy Cheung) of a recent performance shows how they look like when they're put together.
Here it is the sheet music (PDF format) for the box notation of one of the pieces, and here it is the Western music notation of the same piece. The LaTeX source for the box notation and that for the Western music notation are here. (Don't bother to compile it though, a number of other support files are needed.)
Students taking this project are expected to be fluent in programming and have some knowledge of, or willing to learn, the basics of music. Those who have enrolled, or will enrol, in the 6-credit course MUSI2004 University Gamelan 1 or is willing to attend the community gamelan (held every Monday night) are highly preferred.
