AsArcFace: Asymmetric Additive Angular Margin Loss for Fairface Recognition
| العنوان: | AsArcFace: Asymmetric Additive Angular Margin Loss for Fairface Recognition |
|---|---|
| المؤلفون: | Shengyao Zhou, Junfan Luo, Junkun Zhou, Ji Xiang |
| المصدر: | Computer Vision – ECCV 2020 Workshops ISBN: 9783030654139 ECCV Workshops (6) |
| بيانات النشر: | Springer International Publishing, 2020. |
| سنة النشر: | 2020 |
| مصطلحات موضوعية: | Computer science, Margin (machine learning), business.industry, Work (physics), Boundary (topology), Artificial intelligence, Machine learning, computer.software_genre, business, Facial recognition system, computer, Task (project management) |
| الوصف: | Fairface recognition aims to the mitigate the bias between different attributes in face recognition task while maintaining the state-of-art accurancy. It is a challenging task due to high variances between different attributes and unbalancement of data. In this work, we provide an approach to make a fairface recognition by using asymmetric-arc-loss training and multi-step finetuning. First, we train a general model with an asymmetric-arc-loss, and then, we make a mutli-step finetuning to get higher auc and lower bias. Besides, we propose another viewpoint on reducing the bias and use bag of tricks such as reranking, boundary cut and hard-sample model ensembling to improve the performance. Our approach achieved the first place at ECCV 2020 ChaLearn Looking at People Fair Face Recognition Challenge. |
| ردمك: | 978-3-030-65413-9 |
| URL الوصول: | https://explore.openaire.eu/search/publication?articleId=doi_________::1d85fde33980bacb6f50853ce49bd3f3 https://doi.org/10.1007/978-3-030-65414-6_33 |
| حقوق: | CLOSED |
| رقم الأكسشن: | edsair.doi...........1d85fde33980bacb6f50853ce49bd3f3 |
| قاعدة البيانات: | OpenAIRE |
| ردمك: | 9783030654139 |
|---|