Every day, countless images and videos are captured in public spaces, such as streets, parking lots, and transit hubs, as well as during interviews, user studies, UX tests, classroom evaluations, and other scenarios involving participants. This raises crucial questions about personal privacy and data protection. Our demo illustrates how advanced machine learning technologies, such as YOLO (which is very fast but less accurate) and FastRNN (slower but much more accurate), can automatically detect and anonymize sensitive data directly at the point of capture. By quickly identifying traffic participants (pedestrians, cyclists, vehicles, buses, etc.) or human faces, this solution immediately anonymizes or visually highlights detected objects and individuals before the data ever leaves the capturing device.

Privacy preservation is essential in today’s data-driven environment, safeguarding individuals’ identities and ensuring compliance with data protection regulations. This solution demonstrates how privacy protection can be integrated at the earliest stage, on embedded systems or edge devices, ensuring that only anonymized detections are transmitted or stored, thereby significantly reducing privacy risks and building trust.

Explore this demo to experience firsthand how machine learning not only enhances object and face detection but also strengthens privacy protection by embedding anonymization into the data collection process itself.