What is it about?
Player re-identification aims to search players in a database to identify them based on a taken picture. Therefore we extract features from the query image we want to identify and compare them against a gallery database of player image. In the gallery the player identity is known before and based on a similarity measurement we match the query to the identity.
Featured Image
Photo by LOGAN WEAVER | @LGNWVR on Unsplash
Why is it important?
Player re-identification is a important task in order to assess the single player performance. Otherwise there is no clear separation.
Perspectives
We provided a fast and easy to apply approach for re-identification. Overall the performance is very good and does not require any specific data pre-processing or multiple models and ensembling to achieve good results.
Fabian Deuser
Universitat der Bundeswehr Munchen
Read the Original
This page is a summary of: CLIP-ReIdent: Contrastive Training for Player Re-Identification, October 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3552437.3555698.
You can read the full text:
Resources
Contributors
The following have contributed to this page
