Skip Navigation eScience Home | ANU Home | Search FEIT | Search ANU

Faculty of Engineering and Information Technology (FEIT)

Dept. of Computer Science (DCS)

Proposals from Tom Gedeon for:

Honours / Software Engineering / eScience / MComp / MCompHons / PhD / Summer Research Projects

If you are interested in any of the following (or similar) topics and would like to see if you might want to do a project in this area, please e-mail me tom[at]cs.anu.edu.au or come and see me.

1.      Face recognition (supervisor Professor Tom Gedeon )

A number of projects are possible in this area. I am interested in a number of topics, such as image processing for face recognition, HCI research tool to collect facial images, biologically plausible architectures for face recognition, building and modifying computer models of real faces. No previous work on face recognition is necessary. For the image processing topic, some experience with computer graphics would be useful. Example projects include:

1. View morphing: View morphing between two images of an object taken from two different viewpoints produces the illusion of physically moving a virtual camera. We generate compelling 2D transitions between images. Initially we will use the same face with small rotations and produce intermediate images to produce a realistic QuickTimeVR head.
2. Video annotation: Techniques such as PCA can recognise the same face so long as the rotation/expression/illumination is not too different. This can be used to recognise one person throughout a segment of video.
3. 3-D Face model: An average face 3D model can be constructed from a single face image and animated via rotation, a simple face expression model, or illumination. This can then be used to recognise faces in multiple poses, expressions or lighting.

2.      Neural networks theory and applications (supervisor Professor Tom Gedeon )

A number of projects are possible in this area. I am interested in a number of topics, such as extracting rules from neural networks, information retrieval using neural networks, data mining and feature selection, cascade neural network structures, hierarchical neural network structures, and neural network applications. I have published papers in all of these areas with former students so there is plenty of earlier work to build on. No previous experience with neural networks is necessary. Most projects will use the very popular backpropagation neural network training algorithm. Example Projects include:

1. Cascade neural network stuctures can be built automatically without making decisions as to the number of neurons required to solve a problem by adding single neurons. This project would investigate the use of larger chunks such as feature maps as cascade components, which would be useful for recognising images (including faces).
2. Rule extraction: Neural networks can learn complex tasks but face the problem that human beings do not trust them as they can not understand _why_ a particular decision is made. This project focuses on rule extraction for explanation.  

3.      Fuzzy logic theory and applications (supervisor Professor Tom Gedeon )

A number of projects are possible in this area. I am interested in a number of topics, such as automated construction of fuzzy rule bases from data, hierarchical fuzzy systems, fuzzy interpolation, information retrieval using fuzzy logic, universal approximators,  and fuzzy logic applications. I have published papers in all of these areas with former students so there is plenty of earlier work to build on. No previous experience with fuzzy logic is necessary. Example Projects include:

1. Fuzzy document filtering: Web search engines return many documents which are not relevant to queries. Fuzzy techniques to enhance the results from search engines will be very useful. This project will investigate a number of fuzzy techniques for this purpose.
2. Combining uncertainty: Fuzzy logic is a technique which deals with uncertainty well. Sometimes data contains multiple kinds or sources of uncertainty. For example, a pedestrian could report that he saw someone he was quite sure was stealing something. He was not certain and his own reliability is another but different kind of uncertainty. This project is to develop some methods to combine different kinds of uncertainty in intelligence led investigation. (Data from a company specialising in this area will be available.)

4.      Bacterial evolutionary algorithms (supervisor Professor Tom Gedeon)

There are several optimisation methods inspired by natural processes, It has been shown that evolutionary algorithms are efficient tools for solving non-linear, multi-objective and constrained optimizations. The principle is a search for a population of solutions, where tuning is done using mechanisms similar to biological recombination. The operations the bacterial evolutionary algorithm were inspired by microbial evolution. Bacteria can transfer genes to other bacteria, so the parameters are optimised in each bacterium. Example projects:

1. Clustering using BEAs: Clustering is more usually performed by gradient descent algorithms. BEAs should allow more interesting clusters to be formed where there are complex evaluation criteria for cluster validity. Objects clustered could be images (e.g. see projects 1.1 or 4.2), or text (e.g. see projects 2.1, 2.2). Clusters could have predefined target criteria, such as hierarchical structure of a certain depth.
2. Making pictures: If bacteria encoded the components of an abstract computer generated picture, an individual could identify nice and not nice images repeatedly to generate some 'art' which is tuned for their esthetic sense.
3. Scheduling: A previous student has worked on a GA for University exam scheduling. A comparison with bacterial algorithms would be an interesting project.

5.      Traffic simulation and Agents (supervisor Professor Tom Gedeon)

Projects in this area can range from further development of the existing simulation software to research in agent interactions, e.g., in investigating driver characteristics, simulating individual car traffic accidents or modelling of annual road toll, and so on.

6.      Cooperative Robot Communication (supervisor Professor Tom Gedeon)

These robots are individually too weak to move objects, hence they must move objects co-operatively. Only the robot foreman knows what task is to be completed, the other robots over time develop a model of the task and assist the foreman to complete it. Projects in this area can range from development of a simulator for these robots, to wresearch on development of the codebooks the robots use to guess the context of each action.

7.      SMS and On-line Dynamic Surveys (supervisors Dr Ken Taylor (CSIRO) and Professor Tom Gedeon)

VotApedia is an audience response system that doesn't require issuing clickers or need specialist infrastructure. See www.votapedia.com The project will involve design and implement extensions to this controlled but open source project initiated by CSIRO. The technology is in use by Tom this semester so there will be opportunity for immediate trials of new functionality.