In this project, we will focus on monitoring physiological exercise executions. We distinguish two levels of monitoring:

• Level 1 (“quantity”): recognizing that an exercise execution has been done correctly
• Level 2 (“quality”): recognizing that an exercise execution has been done correctly; and if not, incorrectly and providing real-time feedback on what was wrong (e.g., wrong posture, execution speed too fast).

While level 1 can be used to automatically track whether a person follows assigned training plans, level 2 can be used as an assistive system, for example, in a rehabilitation setting.

The challenge in this project is that only off-the-shelve cameras (e.g., a smartphone camera or a web cam) should be used as sensors and that training material is restricted (e.g., it will not be possible to record hundreds of videos as training data showing wrong exercise executions). Given these restrictions, this project should explore what is possible in respect to level 1 and/or level 2. Ideas will be discussed initially with the advisors of this project and a small set of exercises will be agreed upon.

Prerequisites

• Visual computing and/or machine learning background
• A plus: knowledgeable about physio / rehab / sports exercise executions.

How to apply

Please send me an email with the following pieces of information (if you do not answer every point, your application will not be considered):

• When you plan to start the project
• When you plan to finish the project
• A short motivational statement why this project is interesting for you
• A summary why you be a good fit for this project
• Your transcript of records
• How would you approach this topic (max. 1 page)