Naranjo, Ruben (Vicomtech), Aranjuelo, Nerea (Vicomtech), Rodríguez-Moreno, Itsaso (Euskal Herriko Unibertsitatea)

Methodology for Real-Time Privacy Preserving Data Recording on Autonomous Vehicles

Scheduled for presentation during the Regular Session "S12c-Safety and Risk Assessment for Autonomous Driving Systems" (WE-LA-T12), Wednesday, November 19, 2025, 16:40−17:00, Broadbeach 3

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 19, 2025

Abstract

The development and validation of autonomous vehicle technologies necessitates extensive data recording campaigns to capture diverse real-world scenarios. These campaigns inherently involve the collection of privacy-sensitive information, particularly visual data containing individuals in public and private spaces. Ensuring the privacy of these individuals while maintaining the utility of the data for training and testing advanced driver-assistance systems (ADAS) and fully autonomous driving (AD) features is paramount, especially considering regulations like the GDPR. In this paper, we propose a novel two-step methodology for efficient real-time anonymisation onboard vehicles and realistic person synthesis as an offline process. The real-time onboard process anonymises images through masking and collects data to guide the offline synthesis. The offline pipeline synthesises new individuals using collected guidance data to retrieve the usability of the images. Qualitative results show that our approach is on par with state of the art generative anonymisation methods while preserving the real-time constraints of data collection.