A4 Article in conference proceedings
CCTVCV : Computer Vision model/dataset supporting CCTV forensics and privacy applications (2022)

Turtiainen, H., Costin, A., Hämäläinen, T., Lahtinen, T., & Sintonen, L. (2022). CCTVCV : Computer Vision model/dataset supporting CCTV forensics and privacy applications. In TrustCom 2022 : Proceedings of the IEEE 21st International Conference on Trust, Security and Privacy in Computing and Communications (pp. 1219-1226). IEEE. IEEE International Conference on Trust, Security and Privacy in Computing and Communications. https://doi.org/10.1109/trustcom56396.2022.00169

JYU authors or editors

Publication details

All authors or editors: Turtiainen, Hannu; Costin, Andrei; Hämäläinen, Timo; Lahtinen, Tuomo; Sintonen, Lauri

Parent publication: TrustCom 2022 : Proceedings of the IEEE 21st International Conference on Trust, Security and Privacy in Computing and Communications

Place and date of conference: Wuhan, China, 9.-11.12.2022

ISBN: 978-1-6654-9426-7

eISBN: 978-1-6654-9425-0

Journal or series: IEEE International Conference on Trust, Security and Privacy in Computing and Communications

ISSN: 2324-898X

eISSN: 2324-9013

Publication year: 2022

Pages range: 1219-1226

Publisher: IEEE

Publication country: United States

Publication language: English

DOI: https://doi.org/10.1109/trustcom56396.2022.00169

Publication open access: Not open

Publication channel open access: 

Publication is parallel published (JYX): https://jyx.jyu.fi/handle/123456789/86088

Abstract

The increased, widespread, unwarranted, and unaccountable use of Closed-Circuit TeleVision (CCTV) cameras globally has raised concerns about privacy risks for the last several decades. Recent technological advances implemented in CCTV cameras, such as Artificial Intelligence (AI)-based facial recognition and Internet of Things (IoT) connectivity, fuel further concerns among privacy advocates. Machine learning and computer vision automated solutions may prove necessary and efficient to assist CCTV forensics of various types. In this paper, we introduce and release the first and only computer vision models are compatible with Microsoft common object in context (MS COCO) and capable of accurately detecting CCTV and video surveillance cameras in street view, generic images, and video frames. Our best detectors were built using 8,387 images, which were manually reviewed and annotated to contain 10,419 CCTV camera instances, and achieved an accuracy rate of up to 98.7%. This work proves fundamental to a handful of present and future applications that we discuss, such as CCTV forensics, pro-active detection of CCTV cameras, providing CCTV-aware routing, navigation, and geolocation services, and estimating their prevalence and density globally and on geographic boundaries.

Keywords: closed-circuit television; cameras; computer vision; machine learning; artificial intelligence; privacy; data protection; facial recognition (computer science); applications (applying); forensic criminal investigation

Free keywords: CCTV; cameras; computer vision; datasets; machine learning; mapping; object detection; privacy; video surveillance

Contributing organizations

Ministry reporting: Yes

VIRTA submission year: 2023

JUFO rating: 1