AnonymEyes: Lightweight and Affordable Privacy-Enhanced Human Detection via Depth Sensing on Edge Devices
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Privacy concerns in human monitoring systems have become increasingly critical as surveillance technologies proliferate in public and private spaces. Conventional detection solutions rely on RGB-based systems, raising serious privacy concerns due to potential identity disclosure. This paper introduces AnonymEyes, a lightweight, affordable, and privacyenhanced human detection system utilizing Time-of-Flight (ToF) depth cameras with edge computing devices such as the Raspberry Pi. By exclusively leveraging depth data while discarding RGB-based data, AnonymEyes enhances individual anonymity while maintaining high detection performance. The system employs YOLOv8 Nano architecture based framework to process ToF depth frames, achieving real-time detection on resourceconstrained environments. The system was trained and validated on over 1,500 annotated ToF depth frames encompassing diverse poses, clothing, and appearances. The system demonstrates exceptional metrics: 99.90% precision, 100.00% recall, 99.95% F1-score, 99.50% mAP@0.5, and 92.50% mAP@0.5:0.95 across stricter IoU thresholds. AnonymEyes represents a robust and scalable solution for applications requiring accurate, real-time human presence monitoring in privacy-sensitive, resourceconstrained environments.
Fields of Study & Subjects
Advanced Neural Network Applications
Video Surveillance and Tracking Methods
Face recognition and analysis
Publication details
- Journal / source
- International Conference on Artificial Intelligence in Information and Communication (ICAIIC)
- Pages
- 333–338
- Publication date
- 24 February 2026
- Language
- EN