Background

Experts diagnose the severity of knee Osteoarthritis by quantifying key observations (e.g. Bone malformation) in a radiograph and sum them up to describe the overall evaluation in a scale termed as Kellgren-Lawrence score. During the first stage of the project, we have developed semi-automatic segmentation along with geometric and appearance description technique for large scale evaluation. Moreover, an advanced machine learning technique is developed to mimic the original decision making process of the experts based on relative ranking.

Thesis Objective

The goal is to design a crowdsourced study where non-expert user, equipped with a novel computer-aided diagnostic technique combining advanced machine learning and shape analysis, can do a prediction job in quality similar to the experts. The implementation involves a graphical web-based user interface, server-side development as well as integration into the Amazon Mechanical Turk web service.

Pre-requisites

1. Experience in front- and back-end web programming
                - any of the following: C#, Java, Ruby, Perl
                - basic database design

2. It is beneficial to have some game development and (or) crowdsourcing experience.

We Offer

 Insight into the state-of-the-art of medical image analysis in a stimulating research environment. We offer co-supervision for the thesis and close collaboration within the research group over the entire duration of the project.   

Research Group: 
Classifying Lesions of Knee Menisci using nD Views
MA 03/17
Deep Vision for Ecology and Environment
MA 02/17