Explore the use of a light field to get both motion and depth simultaneously by using the plenoptic function and Image Jacobian. The project will consider testing this on Lytro and various robotic systems.
Combine image processing and controls in a smart and adaptive way — something like Lab 3 of METR4202 (2013).
Design and build a robot manipulator using SDM (shape deposition manufacturing) capable of grasping 5 cream filled cookies from a box and placing them into another container. We will start with the Trossen robotic arm and progress to the ABB arm if there is time. The challenge is to make this reliably pickup cookies that have differences in size and thickness (not all cream cookies are absolutely identical)
Remote Access CT imaging Laboratory for clinical skills education and training
Using an initial set of keypoints identified in an image, the project will help track the diving motion of athletes. For Honours: The system will need to be tested against video datasets from NSWIS and demonstrated at NSWIS and AIS.
Make a levitating light
Consider the process of semi-automating the tele-operation process for the diagnosis of melenova nevi using dermoscopy and haptic feedback methods.
Help us help the blind! In this project you will look at making smart tones for a programmable tonal ball for the visually impaired. Co-supervised with Paul Pounds or Hanna Kurniawati Team sports is important to the social and physical development of children with vision impairment. We are developing an assistive sporting device built into a soccer ball that emits tones according to its motion. This allows a player to track and engage with a ball without vision sense. Help develop the next generation of sporting aid! 6 - Levitating Light
Joint with Prof. Udantha Abeyratne – RESEARCH Intensive (RDL*) ] Can we use vision (or other sensing – eg induction or hall effect sensing) to track adult breathing? An initial direction is affine video tracking of features (natural, such as hair, or artificial, such as lines drawn using a wax pencil) on the chest of a patient. By calculating and tracking the affine deformations of the lines between these features, we hope to recover the breathing rate (Hz) and a qualitative measure of breathing strength (nb, this could be extended via regression to prior test cases to be a quantitative measures such as air exchange volume). This project should also consider this in comparison to other proposed methods for this, such as Eulerian Video Magnification [http://people.csail.mit.edu/mrub/vidmag/, tracking using a RGBD camera [Kinect 2], etc.
A Katita is a spring toy. Borrowing from Underactuated Robotics and a Brisslebot, we seek to make it “walk” in a straight line by deferentially driving two vibrator motors (from a mobile phone). This “toy problem” is no “cakewalk” – It combines model predictive control theory and system identification.
Build an ultrasound support using rapid-prototyping (e.g., FDM or Sterolithography) technology.