在人臉辨識這項課題中，我們會先使用深度學習的方法將人臉影像擷取出特徵。而這些特徵如果是來自同一個身份，通常會期望在特徵空間中，這些特徵的距離越近越好。反之，如果是來自不同身份，則距離越遠越好。最近的研究大多透過發展新的損失函數，來同時滿足壓縮類內距離和增加類間差距這兩個目的。這些研究也的確在人臉辨識中達到極高的準確度。
不過，當我們在現實情況下測試時，就會發現表現並不如我們預期，很大一部份的原因在於，現今開源的大型資料集大多內容都是以歐美人的影像為主，亞洲或是非洲臉孔相較之下非常少。而我們實驗主要是用來測試亞洲臉孔，因此有這樣的現象也不是那麼意外了。因次本篇論文就是希望在無法產生大量的亞洲臉孔資料來訓練人臉辨識中擷取特徵模型的情況下，解決以上的問題而設計出的方法。這項方法主要是透過線性判別分析和K-近鄰演算法來實現，也在測試中達到不錯的結果。

In the task of face recognition, we will first use deep learning method to extract features from face images. In the feature space, if these features come from the same identity, or we can say the same person, it is better to get smaller distance between them. Conversely, if they come from different identities, the farther away the better. Most of the recent studies have developed new loss functions for meeting meet two purposes simultaneously. One is compressing the in-tra distance and the other is increasing the in-ter distance. These studies have indeed achieved extremely high accuracy in face recognition.
However, when we test in reality, we will find that the performance is not as we expected. The main factor is that most of the large-scale open dataset are mainly based on European and American images, whether Asian or African face images are very few. And our experiment is mainly used to test Asian faces, so it is not so unexpected to have such a phenomenon. The purpose of this paper is to solve the above problem, in the case that we cannot generate a large number of Asian face data ourselves for training the feature extraction model in face recognition. Our method is realized by linear discriminant analysis and K-nearest neighbor algorithm, and it also achieves good results in the testing process.

[1] K. Zhang, Z. Zhang, Z. Li, Y. Qiao: Joint face detection and alignment using multi-task cascaded convolutional networks. arXiv preprint arXiv:1604.02878, 2016.
[2] Z. Liu, P. Luo, X. Wang, and X. Tang. Deep learning face attributes in the wild. In ICCV, 2015.
[3] J. Deng, J. Guo, and S. Zafeiriou. Arcface: Additive angular margin loss for deep face recognition. arXiv preprint arXiv:1801.07698, 2018.
[4] W. Liu, Y. Wen, Z. Yu, M. Li, Bhiksha Raj, and Le Song. Sphereface: Deep hypersphere embedding for face recognition. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
[5] H. Wang, Y. Wang, Z. Zhou, X. Ji, Z. Li, D. Gong, J. Zhou, and W. Liu. Cosface: Large margin cosine loss for deep face recognition. arXiv preprint arXiv:1801.09414, 2018.
[6] W. Liu, Y. Wen, Z. Yu, and M. Yang. Large-margin softmax loss for convolutional neural networks. In ICML, pages 507–516, 2016.
[7] R. Ranjan, C. D Castillo, and R. Chellappa. L2-constrained softmax loss for discriminative face verification. arXiv preprint arXiv:1703.09507, 2017.
[8] F. Wang, W. Liu, H. Liu, and J. Cheng. Additive margin softmax for face verification. arXiv preprint arXiv:1801.05599, 2018.
[9] F. Wang, X. Xiang, J. Cheng, and A. L Yuille. Normface: l2 hypersphere embedding for face verification. arXiv preprint arXiv:1704.06369, 2017.
[10] K. Zhao, J. Xu, and M.-M. Cheng. Regularface: Deep face recognition via exclusive regularization. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1136–1144, 2019.
[11] H. Liu, X. Zhu, Z. Lei, and S. Z Li. Adaptiveface: Adaptive margin and sampling for face recognition. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 11947–11956, 2019.
[12] X. Zhang, R. Zhao, Y. Qiao, X. Wang, and H. Li. Adacos: Adaptively scaling cosine logits for effectively learning deep face representations. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 10823–10832, 2019.
[13] F. Schroff, D. Kalenichenko, J. Philbin. Facenet: A unified embedding for face recognition and clustering. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 815–823, 2015.
[14] Y. Wen, K. Zhang, Z. Li, and Y. Qiao. A discriminative feature learning approach for deep face recognition. In ECCV, pages 499–515, 2016.
[15] K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
[16] J. Hu, J. Lu, and Y.-P. Tan. Discriminative deep metric learning for face and kinship verification. In TIP, 26(9):4269–4282, 2017
[17] P.-N. Belhumeur, J.-P Hespanha, and D.-J. Kriegman. Eigenfaces vs. fisherfaces: Recognition using class specific linear projection. IEEE Transactions on pattern analysis and machine intelligence, 19(7):711–720, 1997.
[18] K. Q Weinberger and L. K Saul. Distance metric learning for large margin nearest neighbor classification. Journal of Machine Learning Research, 10(Feb):207–244, 2009.
[19] G. B. Huang, M. Ramesh, T. Berg, and E. Learned-Miller. Labeled Faces in the Wild: A database for studying face recognition in unconstrained environments. Technical Report 07-49, University of Massachusetts, Amherst, 2007.
[20] D. Yi, Z. Lei, S. Liao, and S. Z. Li. Learning face representation from scratch. arXiv preprint arXiv:1411.7923, 2014.
[21] M. Wang and W. Deng. Deep face recognition: A survey. arXiv:1804.06655, 2018.
[22] Y. Taigman, M. Yang, M. Ranzato, and L. Wolf. Deepface: Closing the gap to human-level performance in face verification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014.
[23] S. Yang, P. Luo, C. C. Loy, and X. Tang, “WIDER FACE: A Face detection benchmark,” arXiv:1511.06523.
[24] M. Ko ̈stinger, P. Wohlhart, P. M. Roth, and H. Bischof. Annotated facial landmarks in the wild: A large-scale, real-world database for facial landmark localization. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, pages 2144–2151, 2011.