Job Description

Description

Category

Machine Learning, Computer Science

Scope

Master Thesis, 2 students completing 30 credits each
 

Background

A promising research direction of Multiple Object Tracking (MOT) is that of Machine Learning trained trackers.

These methods can contain elements and sub-algorithms within object detection, re-identification, association, transformers and more so it is a very rich and active research field for students to learn and explore and also to gain hands-on machine learning experience.
Some methods are end-to-end from pixel inputs, some are based on object detection boxes, but the common focus in this thesis is that they are trained rather than manually constructed heuristics.

Goal

Answer some or all of the questions:

  • How do trainable trackers perform compared to traditional tracking-by-detection trackers?
  • How do they perform in the video surveilla...

    • Apply for this Position

    Ready to join Axis Communications? Click the button below to submit your application.

    Submit Application