Robust Skin Color Driven Privacy Preserving Face Recognition via Function Secret Sharing
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913419385372672 |
|---|---|
| author | Han, Dong Jiang, Yufan Li, Yong Mendes, Ricardo Denzler, Joachim |
| author_facet | Han, Dong Jiang, Yufan Li, Yong Mendes, Ricardo Denzler, Joachim |
| contents | In this work, we leverage the pure skin color patch from the face image as the additional information to train an auxiliary skin color feature extractor and face recognition model in parallel to improve performance of state-of-the-art (SOTA) privacy-preserving face recognition (PPFR) systems. Our solution is robust against black-box attacking and well-established generative adversarial network (GAN) based image restoration. We analyze the potential risk in previous work, where the proposed cosine similarity computation might directly leak the protected precomputed embedding stored on the server side. We propose a Function Secret Sharing (FSS) based face embedding comparison protocol without any intermediate result leakage. In addition, we show in experiments that the proposed protocol is more efficient compared to the Secret Sharing (SS) based protocol. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_05045 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Robust Skin Color Driven Privacy Preserving Face Recognition via Function Secret Sharing Han, Dong Jiang, Yufan Li, Yong Mendes, Ricardo Denzler, Joachim Computer Vision and Pattern Recognition In this work, we leverage the pure skin color patch from the face image as the additional information to train an auxiliary skin color feature extractor and face recognition model in parallel to improve performance of state-of-the-art (SOTA) privacy-preserving face recognition (PPFR) systems. Our solution is robust against black-box attacking and well-established generative adversarial network (GAN) based image restoration. We analyze the potential risk in previous work, where the proposed cosine similarity computation might directly leak the protected precomputed embedding stored on the server side. We propose a Function Secret Sharing (FSS) based face embedding comparison protocol without any intermediate result leakage. In addition, we show in experiments that the proposed protocol is more efficient compared to the Secret Sharing (SS) based protocol. |
| title | Robust Skin Color Driven Privacy Preserving Face Recognition via Function Secret Sharing |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2407.05045 |