Swimming Pool Occupancy Analysis using Deep Learning on Low Quality Video

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

Abstract

Automatically creating spatio-temporal occupancy analysis of public swimming pools is of great interest, both for administrators to optimize the use of these expensive facilities, and for users to schedule their activities outside peak hours. In this paper we apply current state-of-the-art deep learning methods within human detection on low quality swimming pool video. Furthermore, we propose a method for analyzing the spatio-temporal occupancy of a swimming pool. We show that it is possible to precisely detect swimmers in very challenging conditions by obtaining an AUC of 93.48 % from YOLOv2. An acceptable AUC of 79.29 % was obtained from Tiny-YOLO, which can be implemented on a low-cost embedded system capable of producing results in real-time on site. We expect that the performance of both networks can be improved with more training data.
Original languageEnglish
Title of host publicationACM Multimedia Conference Workshops : First International Workshop on Multimedia Content Analysis in Sports
PublisherAssociation for Computing Machinery
Publication dateOct 2018
Pages67-73
ISBN (Print)978-1-4503-5981-8
DOIs
Publication statusPublished - Oct 2018
EventACM Multimedia Conference Workshops: First International Workshop on Multimedia Content Analysis in Sports - Lotte Hotel Seoul, Seoul, Korea, Republic of
Duration: 22 Oct 201826 Oct 2018

Conference

ConferenceACM Multimedia Conference Workshops
LocationLotte Hotel Seoul
CountryKorea, Republic of
CitySeoul
Period22/10/201826/10/2018

Fingerprint

Swimming pools
Embedded systems
Deep learning
Costs

Cite this

Jensen, M. B., Gade, R., & Moeslund, T. B. (2018). Swimming Pool Occupancy Analysis using Deep Learning on Low Quality Video. In ACM Multimedia Conference Workshops: First International Workshop on Multimedia Content Analysis in Sports (pp. 67-73). Association for Computing Machinery. https://doi.org/10.1145/3265845.3265846
Jensen, Morten Bornø ; Gade, Rikke ; Moeslund, Thomas B. / Swimming Pool Occupancy Analysis using Deep Learning on Low Quality Video. ACM Multimedia Conference Workshops: First International Workshop on Multimedia Content Analysis in Sports. Association for Computing Machinery, 2018. pp. 67-73
@inproceedings{b4791aa8cc1e443fbc58e29f1642f7f7,
title = "Swimming Pool Occupancy Analysis using Deep Learning on Low Quality Video",
abstract = "Automatically creating spatio-temporal occupancy analysis of public swimming pools is of great interest, both for administrators to optimize the use of these expensive facilities, and for users to schedule their activities outside peak hours. In this paper we apply current state-of-the-art deep learning methods within human detection on low quality swimming pool video. Furthermore, we propose a method for analyzing the spatio-temporal occupancy of a swimming pool. We show that it is possible to precisely detect swimmers in very challenging conditions by obtaining an AUC of 93.48 {\%} from YOLOv2. An acceptable AUC of 79.29 {\%} was obtained from Tiny-YOLO, which can be implemented on a low-cost embedded system capable of producing results in real-time on site. We expect that the performance of both networks can be improved with more training data.",
author = "Jensen, {Morten Born{\o}} and Rikke Gade and Moeslund, {Thomas B.}",
year = "2018",
month = "10",
doi = "10.1145/3265845.3265846",
language = "English",
isbn = "978-1-4503-5981-8",
pages = "67--73",
booktitle = "ACM Multimedia Conference Workshops",
publisher = "Association for Computing Machinery",
address = "United States",

}

Jensen, MB, Gade, R & Moeslund, TB 2018, Swimming Pool Occupancy Analysis using Deep Learning on Low Quality Video. in ACM Multimedia Conference Workshops: First International Workshop on Multimedia Content Analysis in Sports. Association for Computing Machinery, pp. 67-73, Seoul, Korea, Republic of, 22/10/2018. https://doi.org/10.1145/3265845.3265846

Swimming Pool Occupancy Analysis using Deep Learning on Low Quality Video. / Jensen, Morten Bornø; Gade, Rikke; Moeslund, Thomas B.

ACM Multimedia Conference Workshops: First International Workshop on Multimedia Content Analysis in Sports. Association for Computing Machinery, 2018. p. 67-73.

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Swimming Pool Occupancy Analysis using Deep Learning on Low Quality Video

AU - Jensen, Morten Bornø

AU - Gade, Rikke

AU - Moeslund, Thomas B.

PY - 2018/10

Y1 - 2018/10

N2 - Automatically creating spatio-temporal occupancy analysis of public swimming pools is of great interest, both for administrators to optimize the use of these expensive facilities, and for users to schedule their activities outside peak hours. In this paper we apply current state-of-the-art deep learning methods within human detection on low quality swimming pool video. Furthermore, we propose a method for analyzing the spatio-temporal occupancy of a swimming pool. We show that it is possible to precisely detect swimmers in very challenging conditions by obtaining an AUC of 93.48 % from YOLOv2. An acceptable AUC of 79.29 % was obtained from Tiny-YOLO, which can be implemented on a low-cost embedded system capable of producing results in real-time on site. We expect that the performance of both networks can be improved with more training data.

AB - Automatically creating spatio-temporal occupancy analysis of public swimming pools is of great interest, both for administrators to optimize the use of these expensive facilities, and for users to schedule their activities outside peak hours. In this paper we apply current state-of-the-art deep learning methods within human detection on low quality swimming pool video. Furthermore, we propose a method for analyzing the spatio-temporal occupancy of a swimming pool. We show that it is possible to precisely detect swimmers in very challenging conditions by obtaining an AUC of 93.48 % from YOLOv2. An acceptable AUC of 79.29 % was obtained from Tiny-YOLO, which can be implemented on a low-cost embedded system capable of producing results in real-time on site. We expect that the performance of both networks can be improved with more training data.

U2 - 10.1145/3265845.3265846

DO - 10.1145/3265845.3265846

M3 - Article in proceeding

SN - 978-1-4503-5981-8

SP - 67

EP - 73

BT - ACM Multimedia Conference Workshops

PB - Association for Computing Machinery

ER -

Jensen MB, Gade R, Moeslund TB. Swimming Pool Occupancy Analysis using Deep Learning on Low Quality Video. In ACM Multimedia Conference Workshops: First International Workshop on Multimedia Content Analysis in Sports. Association for Computing Machinery. 2018. p. 67-73 https://doi.org/10.1145/3265845.3265846