I welcome Honours
students (or Master students) from Computer Science, Electrical Engineering,
Mathematics, IT
(information technology), Optics, Telecommunication, Control and Robotics, Mechatronics, Physcics, Biology,
statistics, etc. to apply our PhD program.
Besides the APA, IPRS, there are ANU and NICTA scholarships
available.
Some sample thesis topics for PhD study:
We have many interested directions on vision research, of both theoretical and application-driven types. Some examples are
(1) Automatic Synthesis of stereoscopic movie from conventional monocular video clips;
(2) Medical image analysis, such as computer aided diagnosis of stereo fundus (retina) images using vision technology;
(3) 3D vision based medical image analysis, image segmentation, vision aided inspection ;
(4) Interactive constraints based scene understanding and modelling from videos. For example reconstructing outdoor scenes, the final model will consist of not only man-made structures, but also natural objects like trees, flowers, and so on.
(5) Low level vision learning for image restoration, image understanding.
(6) Pattern recognition in Pen-based tablet-PC computing environment. ( hand-draw sketch/graph recognizer ) Using a tablet PC doing handwriting input of texts/charts/formulae, or interactively reconstructs some 3d models from an image. 4) Machine Learning in Multi-view-Geometry from exemplar videos.
We welcome students having (but not limited to) one of these backgrounds: information science, electrical/electronics , telecommunication, control systems, computer science, math, physics, engineering, etc.
Constraints-based scenes understanding and modelling from video tapes
In this research, we are going to develop a novel method for building realistic 3D graphic models of very complex scenes from multiple images or video tapes. Most existing methods mainly follow Marr¨s philosophy, which believes that 3D information recovery is the first step of vision perception. They seek methods of constructing fully-automated vision machines. However the results are not satisfactory. Instead, we seek to build a machine vision system that can augment our eyes, rather than replacing our eyes. Therefore, human¨s knowledge will drastically reduce the complexity and increase the performances of a 3D vision modelling system. Here, we mainly want to exploit the observed constraints contained in visual scenes, such as coplanarity, orthogonal and other geometric relations which are present in man-made structures. We believe that some level of interactions between humans and computers will help to realize a more practical and more efficient vision machine.
Learning-based 3D object recognition, localization and tracking
Learning is an essential ability of any intelligent system, needless to say also of our humans. Machine Learning methods should play a more important role in intelligent vision systems. However, how to build such a system that can learn from examples and errors and thus grow its intelligent capability is still an open problem, and also a big challenge as well as a big opportunity for computer vision research. This project aims at developing such a visual learning method, which is applicable for various vision tasks such as image segmentation, 3D object recognition, localization and tracking, as well as 3D reconstruction, etc. Our idea is to augment the method of probabilistic visual learning for object representation (Pentland) by active appearance model techniques. Not only 3D geometries, but also pose/position and kinemic/dynamic properties of moving objects are described in a probabilistic framework. As a result, human¨s a priori knowledge can be incorporated in a natural way and in an early stage of learning, and thus increase the robustness of vision systems.
Pattern recognition and Vision Augmentation in a pen-based (tablet PC) computing environment (Sketch Recognition)
The Tablet PC is a new kind of computer, representing an evolutionary step in the development of the laptop computer used today in mobile computing. It delivers new and easy ways to interact between humans and the computer, and vastly extending the ways in which people will work and enjoy their PCs. On a Tablet PC, users can write/draw directly on its wide screen and save electronic notes in their own handwriting/painting or they can be transformed and utilized in more compact forms through a highly accurate recognition engine. Nowadays there are many character recognition engines available for tablet PCs. The real challenge is to extend such engines to non-text applications. In this research task, we are going to develop several powerful and heterogeneous recognition algorithms that are able to recognize hand-drawn diagrams/sketches/graphs/commercial charts, and mathematical expressions. We can even make a panoramic view or model 3D scenes simply by doodling on the screen. We see the highest potential for success in using a technique called semantic-syntactic structural pattern recognition method which has been proven to be a very powerful recognition method.
Towards True 3D
Recognition: ( Dr. H Li, Prof. R Hartley, Prof. H Burkhardt (Univ.
of Freiburg, Germany ))
Two ultimate goals of computer
vision are reconstruction and recognition. Thanks to the remarkable
achievement of projective vision research in the last decade, the first goal
has gained considerable success. However, the second goal, how to enable
a computer to recognize and understand the real world, is still far from
realized. Instead of using the conventional appearance-based
techniques to recognize, where the appearance (2D image) is a combination
effects of object geometry, material, illuminations, camera properties, and
viewing position etc., which makes the recognition task extremely hard,
our recent work aims at performing recognition task directly on 3D model
obtained from reconstruction (or other scanning techniques). In
other words, we link the reconstruction and recognition procedures in
cascade rather than in parallel. The recognition is solely based on 3D
geometry, so it is true 3D recognition. We propose the space
mapping and harmonic representation techniques as the main mathematical tools,
in hopes that the recognition of objects/scenes would be more
accurate, reliable and robust. Our very early stage trials have received
encouraging results. At the moment, the research is focusing on the study
of invariant representations of 3D geometry.
Graduate students I've supervised and co-supervising/advising:
Fangfang Lu (ANU PhD)
Peter Carr (ANU/NICTA Scholarships)
H.X. Li (CSC Chinese Government Scholarships)
Jae-Hak Kim (NICTA, PhD student, * day-to-day basis),
Yuhang Zhang (RSISE, Research Master student, * day-to-day basis),
Yifan Lu ( RSISE,
PhD student, * day-to-day basis),
Bin Chen, (DCS, Master student, *
day-to-day basis),
Jason Saragih
( RSISE, PhD student, * weekly-meeting),
John Lim (NICTA, PhD student).
Teddy Rasmin, Ph.D student at ANU Vision Group, APAI scholarship.
Thi Bin Chau, Summer Scholar,
Bin Zhu , Summer Scholar, Adelaide University
Wang Jie, Master of Zhejiang U, now doing her PhD study on image processing at U. of Toronto, Canada.
Hui Hua, Master of Zhejiang U, now doing his Ph.D study on signal processing at UIC, USA.
Ziqiang Liu , Master of Zhejiang U and M$R-China Internship, now doing his Ph.D study on statistics for computer vision at UCLA, USA.
Yiming Wu, Master of Zhejiang U , now doing Ph.D study vision at Nanyang U. , Singapore.
Yan Liu , Master of Zhejiang U , now working with Micro-Satellite Institute, Shanghai, China.
Yuguang Hong,Master studying at Zhejiang U.
Jiebing Li , Master studying at Zhejiang U.
Xudan Lou , Master studying at Zhejiang U.
Yong Zhang, Master studying at Zhejiang U.
Yong Li , Master studying at Zhejiang U.
Xin Du, Phd. Candidate, Zhejiang University
Huahua Chen , PhD. student , Zhejiang University