Angular Margin-Mining Softmax Loss for Face Recognition
Face recognition methods have been significantly improved in recent years owing to the advances made in loss functions. Typically, loss functions are designed to enhance the separability power by concentrating on hard samples in mining-based approaches or by increasing the feature margin between dif...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9759281/ |
| _version_ | 1811309339161395200 |
|---|---|
| author | Jwajin Lee Yooseung Wang Sunyoung Cho |
| author_facet | Jwajin Lee Yooseung Wang Sunyoung Cho |
| author_sort | Jwajin Lee |
| collection | DOAJ |
| description | Face recognition methods have been significantly improved in recent years owing to the advances made in loss functions. Typically, loss functions are designed to enhance the separability power by concentrating on hard samples in mining-based approaches or by increasing the feature margin between different classes in margin-based approaches. However, margin-based methods lack the utilization of informative hard sample, and mining-based methods also fail to learn the latent correlations between classes. Moreover, there are no methods that simultaneously consider the effects of hard samples and feature margin through the same shape of feature angular margin. Therefore, this paper introduces the Angular Margin-Mining Softmax (AMM-Softmax) loss function, which adaptively emphasizes hard samples while also increasing the decision margins. The proposed AMM-Softmax loss function introduces a linear angular margin for hard samples, enabling the direct optimization of the geodesic distance margin and maximization of class separability. Furthermore, the proposed AMM-Softmax loss function is computationally efficient and can be easily converged by rapidly switching from the hard samples to easy samples. The results of the extensive experimental analyses conducted on popular benchmarks demonstrate the superiority of the proposed AMM-Softmax loss function over the existing state-of-the-art methods. |
| first_indexed | 2024-04-13T09:40:32Z |
| format | Article |
| id | doaj.art-f1cfafe9b595494db177f7b6f1db75a6 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-13T09:40:32Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-f1cfafe9b595494db177f7b6f1db75a62022-12-22T02:51:56ZengIEEEIEEE Access2169-35362022-01-0110430714308010.1109/ACCESS.2022.31683109759281Angular Margin-Mining Softmax Loss for Face RecognitionJwajin Lee0https://orcid.org/0000-0003-3732-866XYooseung Wang1https://orcid.org/0000-0002-2341-0251Sunyoung Cho2https://orcid.org/0000-0002-6925-6077Defense Artificial Intelligence Center, Agency for Defense Development, Daejeon, Republic of KoreaDefense Artificial Intelligence Center, Agency for Defense Development, Daejeon, Republic of KoreaDefense Artificial Intelligence Center, Agency for Defense Development, Daejeon, Republic of KoreaFace recognition methods have been significantly improved in recent years owing to the advances made in loss functions. Typically, loss functions are designed to enhance the separability power by concentrating on hard samples in mining-based approaches or by increasing the feature margin between different classes in margin-based approaches. However, margin-based methods lack the utilization of informative hard sample, and mining-based methods also fail to learn the latent correlations between classes. Moreover, there are no methods that simultaneously consider the effects of hard samples and feature margin through the same shape of feature angular margin. Therefore, this paper introduces the Angular Margin-Mining Softmax (AMM-Softmax) loss function, which adaptively emphasizes hard samples while also increasing the decision margins. The proposed AMM-Softmax loss function introduces a linear angular margin for hard samples, enabling the direct optimization of the geodesic distance margin and maximization of class separability. Furthermore, the proposed AMM-Softmax loss function is computationally efficient and can be easily converged by rapidly switching from the hard samples to easy samples. The results of the extensive experimental analyses conducted on popular benchmarks demonstrate the superiority of the proposed AMM-Softmax loss function over the existing state-of-the-art methods.https://ieeexplore.ieee.org/document/9759281/Deep convolutional neural networkface recognitionmetric-learning and softmax loss function |
| spellingShingle | Jwajin Lee Yooseung Wang Sunyoung Cho Angular Margin-Mining Softmax Loss for Face Recognition IEEE Access Deep convolutional neural network face recognition metric-learning and softmax loss function |
| title | Angular Margin-Mining Softmax Loss for Face Recognition |
| title_full | Angular Margin-Mining Softmax Loss for Face Recognition |
| title_fullStr | Angular Margin-Mining Softmax Loss for Face Recognition |
| title_full_unstemmed | Angular Margin-Mining Softmax Loss for Face Recognition |
| title_short | Angular Margin-Mining Softmax Loss for Face Recognition |
| title_sort | angular margin mining softmax loss for face recognition |
| topic | Deep convolutional neural network face recognition metric-learning and softmax loss function |
| url | https://ieeexplore.ieee.org/document/9759281/ |
| work_keys_str_mv | AT jwajinlee angularmarginminingsoftmaxlossforfacerecognition AT yooseungwang angularmarginminingsoftmaxlossforfacerecognition AT sunyoungcho angularmarginminingsoftmaxlossforfacerecognition |