Master Assignment
[B] Develop an object detection algorithm to restore ‘vision’ for the visually impaired
Type: Master BME/EE/CS/HMI
Location: (Optional)
Period: Sept, 2021 - TBD
Student: (Unassigned)
If you are interested please contact :
Description:
Do you want to combine image processing and computer vision with medical innovation? Then this is the project for you! TIM Solutions, a MedTech company, is developing a device for blind people to safely live their daily life.
Background information:
Blind and partially sighted people are forced to create their own ways of scanning their surroundings and gaining a sense of direction. Many ideas have been devised to support this group. Yet, to date, nothing has reached the market that is consistently used alongside the conventional white cane and/or guide dog. TIM Solutions has developed a prototype vest in which multiple vibrating actuators, located on the back in a specific pattern, convert obstacles captured by an AI depth camera into vibrotactile information.
Assignment description:
The main part of the assignment is focused on developing an algorithm, based on open-source available software, to detect obstacles with an AI depth-camera and translate this information into understandable vibrotactile information. To be able to develop an algorithm that matches demand, it is important to know in advance what the visually impaired actually need. How quickly should they be warned? How frequent, powerful, and intense should the feedback be? Etcetera. So, the assignment also involves getting familiar with obstacles commonly encountered by the target group (e.g. stairs, branches, bikes and bins for instance) and identifying key elements to enable their detection in a reliable way. At the end of the thesis, the prototype (algorithm and vest) will be tested by practical experiments with the target group.
Preferred study program:
Biomedical Engineering or another program with knowledge about image processing and computer vision