Automatic detection of abnormal video events in sport videos

Automatic detection of abnormal video events in sport videos

Use machine learning and video processing techniques to automatically find events in sports videos. For example, in football, some "easy" ones are goals, but others like tackles are harder.

Sports events are currently manually tagged by human operators in a tedious process. To enable fast and efficient distribution of soccer event (can be video in general), we have developed a system where an operator just press a button for the type of event (like a goal) and publish. However, such systems or manual processes usually focus on traditional events like goals, penalties, cards, etc. whereas there are a number of event types that are of interest to the supporters and fans like tackles, mistakes, huge misses, etc.

The idea of this task is to use machine learning and video processing in order to find such events. A site like contains the traditional manually tagged events like goals, cards, free-kicks, etc. However, fans often like other events not part of the traditional statistics.


A system (or machine learning model) that can assist the annotation (production) of video events in a video streaming service running for the Norwegian and Swedish elite leagues in soccer. The goal is to find “abnormal” events like tackles or other controversial events.

Learning outcome

Video processing, machine learning, real-world implementation and experiments.


  • Python programming
  • Knowledge about deep learning and video processing is an advantage


  • Pål Halvorsen
  • Michael Riegler

Collaboration partners

  • ForzaSys AS

Associated contacts

Pål Halvorsen

ProfessorHead of DepartmentChief Research Scientist/Research Professor

Michael Riegler

ProfessorChief Research Scientist/Research Professor