¡¡

Return to  Home Page

¡¡Scholarships are available for both ANU supported summer internship students and NICTA supported summer students.

Summer student scholarships (for undergraduates and master students.)

Come to the ANU College of Engineering and Computer Science and enjoy 8-12 weeks of study on an exciting research project. Take advantage of the chance to work with some of the world¡¯s leading engineers and computer scientists at ANU and NICTA and you will also receive:

  • A weekly tax-free allowance of $135
  • Return travel to Canberra

Closing date for applications: 31 August every year !

 

How to apply

 

Offer is open to Undergraduate and Honours students who are currently enrolled at any university in Australia or in New Zealand. Under some circumstances, other students who have been accepted into an ANU higher degree starting the following year, may be eligible to apply.

 

>>>    New Projects are updating frequently on the CECS website, please check there¡­>>>>

                  Summer Projects for Year-2009

***********************************************

(1)  Where am I ?--- Computer Vision-based  localization system (a camera-phone based GPS)

Supervisor :  Dr  Hongdong LI.   Email:  Hongdong.li@anu.edu.au , html:  www.rsise.anu.edu.au/~hongdong

 

¡®Location¡¯ is a significant type of information that we are using for every day life.   The extremely successful stories of GPS system and Google-Map are evidences of its significance.   This project aims at developing a new computer vision-based technique for location recognition, using a common mobile camera-phone.     

This project borrows idea from a recent ICCV student contest at http://research.microsoft.com/iccv2005/Contest/  where a detailed project description can be found (ICCV 2005).

However, for this summer project we will develop a much simplified prototype system based on ANU¡¯s campus map.   The main technical components that will be implemented include image feature detecting, image matching, and image database retrieval.   

The student is expected to have some experiences with Matlab or C/C++ programming, have basic knowledge of linear algebra, and feel exciting in doing image processing and computer vision research.   The student will be working closely on a day-to-day basis with the supervisors and PhD students.

******************************************

(2)       Remove camera shake (i.e., motion blur) from a single image (photo). 

Supervisor :  Dr  Hongdong LI.   Email:  Hongdong.li@anu.edu.au , html:  www.rsise.anu.edu.au/~hongdong

 

Unless you are using an expensive professional digital camera with VR (vibration reduction) functionality, camera shake during exposure often lead to objectionable image blur and ruins many photographs. 

 Conventional image post-processing methods typically assume overly simplified forms for the camera motion path during camera shake.   

In this project we will explore and experiment on a newly-introduced method to remove the effects of camera shake (or motion blur) from a single seriously blurred image (ref: [1][2]).  

[1] Removing camera shake from a single image, Fergus  et al, SIGGRAPH 20006.

[2] Blind Motion Deblurring Using Image Statistics, Levin, NIPS 2006.

The student would be expected to have knowledge of calculus and linear algebra, Matlab or C/C++ programming experience.    Having basic knowledge of signal processing or linear system or control will be beneficial.   The student will work closely on a daily basis with the supervisor and postgraduate students.

 

******************************************************************

 

(3)              Interactive matching and camera tracking based on dynamic programming  

 

Supervisor :  Dr  Hongdong LI.   Email:  Hongdong.li@anu.edu.au , html:  www.rsise.anu.edu.au/~hongdong

 

Feature matching and camera tracking are the process of extracting long and accurate tracks of 3D features observed in 2D video.   Points of interest are indicated with a single mouse click in one frame of the video, and the desired output of the tracker is the location of the point¡¯s 2D projection in every frame of the sequence.  This technique is very useful in producing special video effect in a movie or generating realistic video scene in video games.

 

In this project we will study a new algorithm based on dynamic programming ([1]) .  This new algorithm works very fast and reliably. It is based on the k-d trees and dynamic programming ¡ªboth are well-known techniques in computer science and software engineering.   

 

[1] Interactive Feature Tracking using K-D Trees and Dynamic Programming, Buchanan et.al.,  CVPR 2006.

 

***********************************************

 

 

(4)             Selective  photo re-blurring using depth-map estimation from a single image

 

Supervisor :  Dr  Hongdong LI.   Email:  Hongdong.li@anu.edu.au , html:  www.rsise.anu.edu.au/~hongdong

 

In professional portrait photograph it is often desired to have a sharp foreground with a blurred background.  This is traditionally achieved by using a profession-grade DSLR camera with big-aperture lenses.   However, many inexpensive point-and shoot digital cameras are often unable to implement such a selective-blurring effect, because of their relatively small lenses. 

 

In this project, we will study and implement a new method for digital photo post-processing to simulate such a selective blurring effect from a photo obtained by a point-and-shot camera.  At the hart of the method is a computer vision algorithm that estimates a raw depth-map estimation from a single image. 

 

Ref [1]:  Defocus Magnification, Eurographics 2007.  

The student would be expected to have knowledge of calculus and linear algebra, Matlab or C/C++ programming experience.    Having basic knowledge of image processing will be beneficial.   The student will work closely on a daily basis with the supervisor and postgraduate students.

 

(5)        GPU-based efficient computations of multi-view vision geometry.

 

 

*****************************************************

 

 

 

Past  Summer Student Scholarship/Internship Research Projects (Since 2003)

NEW SUMMER INTERNSHIP PROJECTS 

1. Natural Image Matting:  Bayesian approach vs. direct solution.

Supervisor :  Hongdong LI Dr.  Email:  Hongdong.li@anu.edu.au

Image matting (also known as blue screen technique) --- extracting foreground element from a background image --- has been a special effect technique in Hollywood industry for decades.   However, not until very recently have many computer vision researchers been involved in this field.   Using advanced computer vision-based matting algorithms (for example,  Bayesian inference), now a film-maker can easily insert a human character into a different scene without annoying visual artefacts.   An excellent website of a popular image matting algorithm is located at  http://grail.cs.washington.edu/projects/digital-matting/image-matting/.   

In this project, the student will study a new method of image matting.  This method is a natural extension of a direct solution algorithm proposed by Levin at CVPR-2006.   Comparison experiments between this new algorithm and conventional Bayesian approach will be conducted.   New insights and more effective solution are expected to reach at the completion of this project. 

The student is expected to have programming experiences with Matlab or C/C++,  have basic knowledge of image processing/computer vision, have a good grasp of linear algebra or linear system theories, and feel fun in doing image and graphics.  The student will be working closely on a day-to-day basis with the supervisor.

2.  Multi-view Vision 3D Reconstruction:  Bundle adjustment vs. L-infinity programming

Supervisor :  Hongdong LI Dr.  Email:  Hongdong.li@anu.edu.au

3D Reconstruction is one of the most central and classical tasks in computer vision research.  While many conventional algorithms such as Bundle-adjustment enjoys many successes in small-size toy problems, these conventional algorithms, when facing large scale real world applications, often encounter unexpected difficulties.   L-infinity Programming is a new concept (as well as a new computational framework) that was proposed recently by our group (RH and FS), which is proven to be very promising to many hard vision problems, particularly the multi-view reconstruction/triangulation problem. 

The major tasks involved in this summer research project include the implementing and quantitatively comparing of both Bundle-adjustment and L-infinity algorithms.   The student will learn how to build accurate 3D model from several input images, will begin to appreciate a principled way of conducting creative scientific research. 

The student is expected to have programming experiences with Matlab or C/C++,  have basic knowledge of image processing/computer vision, have a good grasp of linear algebra or linear system theories, and feel fun in doing image processing and vision geometry.  The student will be working closely on a day-to-day basis with the supervisor and a PhD student.

¡¡

3.  Where am I ?---Vision-based localization algorithm research and experiment

Supervisor :  Hongdong LI Dr.  Email:  Hongdong.li@anu.edu.au

¡¡

¡®Location¡¯ is a very significant type of information that we use for every day life.  The extremely successful stories of GPS system and Google-Map are evidences of such significance.   This project aims at developing a new computer vision-based technique for location recognition.   The basic concept is that, wherever you lost in a foreign city, take a photo of a landmark-type building (i.e, not a small shed) using your camera phone, and send it to a central server.   The server will then tell you in front of which building you are standing. Or, hopefully it may also be able to tell your current orientation.   

 

This project borrows idea from a recent ICCV student contest:  ¡°Where am I ?¡± @ http://research.microsoft.com/iccv2005/Contest/  where a detailed project description  can be found.  However, for this summer project we will aim at developing a concept-proving prototype system. The main technical components that will be implemented include SIFT feature detecting, RANSAC matching, and database retrieval.   

 

The student is expected to have programming experiences with C/C++ (Gnu or MS Visual-C++) ,  have basic knowledge of image processing/computer vision, have a good grasp of linear algebra or linear system theories, and feel fun in doing image processing and vision.  The student will be working closely on a day-to-day basis with the supervisors and PhD students.

¡¡

4.  Stereo Panorama: Build  3D stereo panoramic image from monocular sequence

5.  Articulated human body geometry recovery:  factorization method

¡¡

My projects for  year (2004-2005) were:

3D Objects Representation and Retrieval (CVIP-1)

Location: Department of Systems Engineering, RSISE, ANU

Supervisor: Dr Hongdong Li (hongdong.li ¡°at¡± anu.edu.au)

As it is easy to build 3D models from real scene/object today, more and more 3D models are available over the Internet. The growing size of available 3D objects makes 3D object retrieval very important (i.e., a 3D-search-engine like a 3D-Google!). This project focus on exploring some new efficient 3D shape representation techniques such as 3D-Fourier-Analysis and Fourier-On-Sphere ) to better represent 3D geometry and get better retrieval performances. The major tasks involved in this summer research project include the analysis of the properties of these new 3D shape descriptors, and testing the 3D retrieval/classification capability. The student would be expected to have knowledge of signal processing and linear systems theory, linear algebra and calculus, with some programming skills. The student will work closely on a day-to-day basis with supervisor and postgraduate students.

Virtual Stereoscopic 3D Video Synthesis (CVIP-2)

Location: Department of Systems Engineering, RSISE, ANU

Supervisor: Dr Hongdong Li (hongdong.li ¡°at¡± anu.edu.au)

Enjoy impressive 3D effects on your PC without wearing laborious special eyeglasses? The aim of this project is to provide an easier and cheaper way to convert a standard 2D video/image into true holographic-like 3D video/images, and display it on an eyeglasses-free stereo monitor. We are going to use some state-of-the-art computer vision technologies to perform such conversion, and the generated stereo video will then be displaying on a lenticular screen PC monitor in real time. The major tasks of this summer project are to investigate some computer vision techniques for novel-view image synthesis, as well as to implement near real-time stereo video playback software using C++ language. The student would be expected to have some programming experiences with C/C++ or Matlab, have basic knowledge of image processing/computer vision, and linear algebra. The scholar will work closely on a day-to-day basis with the supervisor and a PhD student.


Digital Inpainting of Images/Photos by using Partial Differential Equations (CVIP-3)

Location: Department of Systems Engineering, RSISE, ANU
Supervisor: Dr Hongdong Li (hongdong.li ¡°at¡± anu.edu.au)
Where ancient paint has flaked away, or old photos have scratches, a restoration effort may need to be taken to fill in the ragged scars¡ªa practice known as inpainting. This process is usually highly time-consuming and highly subjective. Researchers are now developing computer techniques to automate such image restoration. This project aims at investigating the most powerful image inpainting approach based on Partial Differential Equations (PDEs) which was suggested by Bertalmio et.al. at University of Minnesota. In this approach, the image-restoration procedure has been modelled as a diffusion process. What we will do in this project is simply to simulate this process by solving a special PDE equation by using Matlab toolbox.

The student would be expected to have knowledge of calculus (differential equations), Matlab programming/image processing. The student will work closely on a daily basis with the supervisor and postgraduate students.

Return to  Home Page

¡¡