Capstone Project

Back to listing
Group 2016-10 Status completed
Title Xbox Kinect-based analysis of jumping to prevent knee injury in female athletes
Supervisor Dr. Thomas Fevens, Dr. Hassan Rivaz, Dr. Paul Martineau (McGill)
Description Goal: Female soccer players are 8 times more at risk of tearing their ACL compared to male players. Part of this is due to the jumping skills and dynamics of jumping; male players jump such that their ACL is less strained. In this project, Microsoft Kinect will be used to compare jumping dynamics of male and female subjects, and give a score to female jumpers based on how similar they jump to male jumpers. Xbox Kinect includes two infrared (IR) cameras that provide depth map of the scene, in addition to an RGB camera that provides video stream. It then segments the scene into body parts using a machine learning algorithm, as shown in the figure below. The segmentation of different body parts can then be used to infer skeleton structure, as shown in the figure below.The aim of this project is to retrieve skeleton motion from Kinect, and perform motion analysis to classify a subjects’ jumping skills. The project can be divided into 3 phases: Phase 1. Operating Kinect from a PC and getting familiar with Kinect SDK (1 month) Phase 2. Getting motion and skeleton data from Kinect (1 month) Phase 3. (4 months) a) Collecting data of male and female subjects while they jump b) Developing classification techniques for automatically assigning a score to jumping. Phase 4. (2 months) Investigating whether new smartphones that come with 2 back cameras can be used instead of Kinect. There are numerous applications for this project. One application is in preventing female soccer players in injuring their ACL. Another application is in musculoskeletal (MSK) surgery, where it is critical to monitor patients who undergo surgery to ensure that they are on the road to recovery. Attending regular physiotherapy sessions is one solution, but is not feasible for all patients especially in remote areas. Gate analysis laboratories are also extremely expensive and are only available in few centres in Canada. An automatic quality assessment technique allows convenient monitoring of these patients at minimum cost.
Student Requirement Image processing Numerical analysis C++ Programming
Tools Kinect SDK Kinect camera
Number of Students 3 or 4
Students Javier Fajardo,Nicole Cappadocia-Assaly,Thinesh Thuraisingam
Comments:
Links: