Automated Privacy-Preserving Video Processing through Anonymized 3D Scene Reconstruction

Why anonymization?

During these last years, the presence of video cameras in public spaces has widely increased, due to the cost reduction and quality increase of technologies such as CCTV and surveillance systems. Their ubiquity has raised a common concern about privacy, and about how individuals could keep control over their own personal information in these scenarios.

As a result, a general consensus has been reached: a trade-off between security and privacy is needed. In Europe, the GDPR (or General Data Protection Regulation) has been approved, with the goal of protecting citizens’ right to privacy. These regulations forbid private entities and institutions to store these kinds of video recordings in the long term, unless it can be guaranteed that the videos have been completely anonymized.

Is there a way to fully remove all personal data from a video, while preserving the intelligibility of the scene? How could that be done?

An example of this type of institution is Paaspop, a music festival taking place in the Netherlands. Its organizers would like to store their surveillance recordings from the surroundings of the festival, to be able to analyze them in the long term with the goal of improving their logistics, in aspects like crowd control or access control to the festival area. However, GDPR prevents them from doing so without ensuring anonymity within the video recordings.

Do you want to be part of our research center?

Don’t hesitate to reach out! Or apply directly to one of our assignments.