SPL-BEV: Soccer Player Localization and Birds-Eye-View Estimation
Persson, Ivar LU ; Ardö, Håkan LU and Nilsson, Mikael LU
(2025)
In Lecture Notes in Computer Science
15621.
p.114-123
- Abstract
- In this work we present SPL-BEV, a method to localize soccer players on a pitch from a monocular RGB camera. SPL-BEV features a network with few parameters that does not need to make any explicit object detection before localization is made. With SPL-BEV we show increased performance on the Spiideo SoccerNet SynLoc dataset compared to the best provided baseline result. The SPL-BEV system samples features from the U-Net feature space using bi-linear interpolation, guided by camera calibration, to generate features at grid points across multiple planes in a 3D world coordinate system. This forms a voxel feature space, which is then processed into grid cell detections on the ground plane, with final location refinement through x/y correction.... (More)
- In this work we present SPL-BEV, a method to localize soccer players on a pitch from a monocular RGB camera. SPL-BEV features a network with few parameters that does not need to make any explicit object detection before localization is made. With SPL-BEV we show increased performance on the Spiideo SoccerNet SynLoc dataset compared to the best provided baseline result. The SPL-BEV system samples features from the U-Net feature space using bi-linear interpolation, guided by camera calibration, to generate features at grid points across multiple planes in a 3D world coordinate system. This forms a voxel feature space, which is then processed into grid cell detections on the ground plane, with final location refinement through x/y correction. The code for SPL-BEV is also published open source on GitHub. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/0f1d3b95-f109-4fc8-9ff3-e2baa2ee8d83
- author
- Persson, Ivar
LU
; Ardö, Håkan
LU
and Nilsson, Mikael
LU
- organization
-
- Computer Vision and Machine Learning (research group)
- LU Profile Area: Natural and Artificial Cognition
- ELLIIT: the Linköping-Lund initiative on IT and mobile communication
- Machine Learning, Systems and Control (master)
- LTH Profile Area: Engineering Health
- Mathematical Imaging Group (research group)
- publishing date
- 2025-10-17
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- Computer Analysis of Images and Patterns : 21st International Conference, CAIP 2025, Las Palmas de Gran Canaria, Spain, September 22–25, 2025, Proceedings, Part I - 21st International Conference, CAIP 2025, Las Palmas de Gran Canaria, Spain, September 22–25, 2025, Proceedings, Part I
- series title
- Lecture Notes in Computer Science
- volume
- 15621
- pages
- 10 pages
- publisher
- Springer Nature
- external identifiers
-
- scopus:105017368174
- ISSN
- 1611-3349
- ISBN
- 978-3-032-04968-1
- DOI
- 10.1007/978-3-032-04968-1_10
- language
- English
- LU publication?
- yes
- id
- 0f1d3b95-f109-4fc8-9ff3-e2baa2ee8d83
- date added to LUP
- 2025-10-03 13:53:29
- date last changed
- 2026-01-23 11:11:28
@inproceedings{0f1d3b95-f109-4fc8-9ff3-e2baa2ee8d83,
abstract = {{In this work we present SPL-BEV, a method to localize soccer players on a pitch from a monocular RGB camera. SPL-BEV features a network with few parameters that does not need to make any explicit object detection before localization is made. With SPL-BEV we show increased performance on the Spiideo SoccerNet SynLoc dataset compared to the best provided baseline result. The SPL-BEV system samples features from the U-Net feature space using bi-linear interpolation, guided by camera calibration, to generate features at grid points across multiple planes in a 3D world coordinate system. This forms a voxel feature space, which is then processed into grid cell detections on the ground plane, with final location refinement through x/y correction. The code for SPL-BEV is also published open source on GitHub.}},
author = {{Persson, Ivar and Ardö, Håkan and Nilsson, Mikael}},
booktitle = {{Computer Analysis of Images and Patterns : 21st International Conference, CAIP 2025, Las Palmas de Gran Canaria, Spain, September 22–25, 2025, Proceedings, Part I}},
isbn = {{978-3-032-04968-1}},
issn = {{1611-3349}},
language = {{eng}},
month = {{10}},
pages = {{114--123}},
publisher = {{Springer Nature}},
series = {{Lecture Notes in Computer Science}},
title = {{SPL-BEV: Soccer Player Localization and Birds-Eye-View Estimation}},
url = {{http://dx.doi.org/10.1007/978-3-032-04968-1_10}},
doi = {{10.1007/978-3-032-04968-1_10}},
volume = {{15621}},
year = {{2025}},
}