Deep learning for particle tracking
Typ
Examensarbete för kandidatexamen
Program
Publicerad
2020
Författare
Lundell, Adrian
Tonderski, David
Meisingseth, Fredrik
Kristiansson, Dennis
Modellbyggare
Tidskriftstitel
ISSN
Volymtitel
Utgivare
Sammanfattning
The use of machine learning for classifcation has in recent years spread into a wide range of disciplines,
amongst them the detection of particles for particle tracking on microscopy data. We
modifed the Python package DeepTrack, which makes use of deep learning to detect particles,
creating a package called U-Track. By using a new network architecture based on a U-Net, better
performance and higher computational effciency than DeepTrack was achieved on images with
multiple particles. Furthermore, functionality to track detected particles over series of frames was
developed. The application of U-Track on experimental data from two-dimensional
ow nanometry
produced tracks consistent with theory, as well as tracking larger quantities of particles over
longer periods of time compared to a digital filter based benchmark algorithm.