Automatic generation of sport highlights

A large number of sport videos exists, and generating automatich highlights or summaries is a detious and time consuming task. In this thesis, we aim to automatically generate highlight summaries from soccer games using tools like machine learning, computer vision and video analysis.
Master

Together with Forzasys, SimulaMet is researching video systems. We have a system that is currently running live in the top leagues in Norway (Eliteserien and OBOS-ligaen) and Sweden (Allsvenskan, Superettan and the Swedish Futsal league). It is an integrated solution supporting operations from video annotation (like goals) to delivery to the client device. Thus, all video from all games are available together with all metadata of official events.

In this respect, this master project will reseach how to make automatic highlights and summaries that can be posted on newspapters, used in sport channels, etc.

Goal

Create a video summarization tool for soccer videos.

Learning outcome

Video management, integrated systems, computer vision, machine learning, video content analysis.

Qualifications

Good programming skills.

Supervisors

  • Pål Halvorsen
  • Michael Riegler

Collaboration partners

ForzaSys AS

References

forzasys.com/forzify.html

highlights.eliteserien.no

highlights.allsvenskan.se

Contact person