Multi-target tracking using event-camera data for microscopy

Event-cameras ( represent a novel sensing modality for fast detection, imaging, and tracking. They have potential applications in surveillance, automotive, and robotics and microscopy. In contrast to classical cameras, such detectors only record changes in light intensity and do not produce standard images. Thus, they require existing (restoration, detection and tracking, SLAM) methods to be adapted or new, tailored methods to be developed.

In this project, a student will be working on implementing state-of-the art detection and tracking algorithms for the analysis of populations of microscopic objects. Using the extremely fast frame rate (100k per second) of event-cameras, it is now possible to mesure the position, size/deformation and velocity of extremely fast objects, and this project will concentrate on the analysis particles and pathogens in circulating fluids.

This project is computational and does not involve lab-based work nor data collection since data data already available. However, additional data can be collected if needed in light of the findings during this project.
The implementation will primarily be achieved using Matlab, using existing toolboxes but other programming languages can be considered.

Required skills:
-Signal/Image processing
- Experience with Matlab
- Experience C/C++ or python is a plus

Supervisor name: 
Yoann Altmann
Supervisor and Deputy email addresses:
Project location: 
Earl Mountbatten Building, Riccarton
Deputy name: 
Yvan Petillot