This artifact provides the complete three-tier reference implementation, dataset subset, and evaluation code necessary to reproduce the experiments and results reported in the paper "Now You See Me, Now You Don’t: Consent-Driven Privacy for Smart Glasses" [1]. Tier 1 implements on-device face blurring and encryption to prevent exposure of raw facial pixels. Tier 2 provides wearers with an optional landmark-driven synthetic face replacement feature to preserve wearer utility. Finally, Tier 3 allows wearers to utilize a consent-mediated restoration protocol that returns original faces only after explicit bystander authorization. The repository includes the end-to-end capture pipeline, companion synthetic-replacement code, restoration and TTP simulation modules, evaluation notebooks, and a sampled dataset of 16,500 annotated frames. The artifact (including an extensive README.md file) can be publicly accessed using the following GitHub repository [2].https://github.com/SYSNET-LUMS/SmartGlassesPrivacy https://github.com/SYSNET-LUMS/SmartGlassesPrivacy
IEEE International Conference on Pervasive Computing and Communications Workshop (PERCOM-W)
2026-03-20
2026-07-21