Black-Box Face Recovery from Identity Features
Abstract
In this work, we present a novel algorithm based on an it-erative sampling of random Gaussian blobs for black-box face recovery, given only an output feature vector of deep face recognition systems. We attack the state-of-the-art face recognition system (ArcFace) to test our algorithm. Another network with different architecture (FaceNet) is used as an independent critic showing that the target person can be identified with the reconstructed image even with no access to the attacked model. Furthermore, our algorithm requires a significantly less number of queries compared to the state-of-the-art solution.
Cite
Text
Razzhigaev et al. "Black-Box Face Recovery from Identity Features." European Conference on Computer Vision Workshops, 2020. doi:10.1007/978-3-030-68238-5_34Markdown
[Razzhigaev et al. "Black-Box Face Recovery from Identity Features." European Conference on Computer Vision Workshops, 2020.](https://mlanthology.org/eccvw/2020/razzhigaev2020eccvw-blackbox/) doi:10.1007/978-3-030-68238-5_34BibTeX
@inproceedings{razzhigaev2020eccvw-blackbox,
title = {{Black-Box Face Recovery from Identity Features}},
author = {Razzhigaev, Anton and Kireev, Klim and Kaziakhmedov, Edgar and Tursynbek, Nurislam and Petiushko, Aleksandr},
booktitle = {European Conference on Computer Vision Workshops},
year = {2020},
pages = {462-475},
doi = {10.1007/978-3-030-68238-5_34},
url = {https://mlanthology.org/eccvw/2020/razzhigaev2020eccvw-blackbox/}
}