We introduce one extra hyperparameter (subcenter number loss_K) to ArcFace to relax the intra-class compactness constraint. In our experiments, we find loss_K=3 can achieve a good balance between accuracy and robustness.
The training process of Subcenter ArcFace is almost same as ArcFace The increased GPU memory consumption can be easily alleviated by our parallel framework.
- MS1MV0 (The noise rate is around 50%), download link (baidu drive, code
8ql0) (dropbox)
1). Train Sub-center ArcFace (loss_K=3) on MS1MV0.
2). Drop non-dominant subcenters and high-confident noisy data (>75 degrees).
python drop.py --data <ms1mv0-path> --model <step-1-pretrained-model> --threshold 75 --k 3 --output <ms1mv0-drop75-path>
3). Train ArcFace on the new MS1MV0-Drop75 dataset.
baidu drive code 3jsh. gdrive
If you find Sub-center ArcFace useful in your research, please consider to cite the following related papers:
@inproceedings{deng2020subcenter,
title={Sub-center ArcFace: Boosting Face Recognition by Large-scale Noisy Web Faces},
author={Deng, Jiankang and Guo, Jia and Liu, Tongliang and Gong, Mingming and Zafeiriou, Stefanos},
booktitle={Proceedings of the IEEE Conference on European Conference on Computer Vision},
year={2020}
}