FaceNet: A unified embedding for face recognition and clustering
Open accessThe DOI matches its Crossref record; 23 of the 23 references it cites were found in a registry or an index.
Abstract
Despite significant recent advances in the field of face recognition [10, 14, 15, 17], implementing face verification and recognition efficiently at scale presents serious challenges to current approaches. In this paper we present a system, called FaceNet, that directly learns a mapping from face images to a compact Euclidean space where distances directly correspond to a measure offace similarity. Once this space has been produced, tasks such as face recognition, verification and clustering can be easily implemented using standard techniques with FaceNet embeddings asfeature vectors. Our method uses a deep convolutional network trained to directly optimize the embedding itself, rather than an intermediate bottleneck layer as in previous deep learning approaches. To train, we use triplets of roughly aligned matching / non-matching face patches generated using a novel online triplet mining method. The benefit of our approach is much greater representational efficiency: we achieve state-of-the-artface recognition performance using only 128-bytes perface. On the widely used Labeled Faces in the Wild (LFW) dataset, our system achieves a new record accuracy of 99.63%. On YouTube Faces DB it achieves 95.12%. Our system cuts the error rate in comparison to the best published result [15] by 30% on both datasets.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Current face recognition systems [8,24,43], already reach very high levels of accuracy in public non-forensic benchmarks, and it is expected that in the coming years they will keep improving.”
From Calibration of score based likelihood ratio estimation in automated forensic facial image comparison (2022) · cites this paper as [43]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“An increasing number of UDA techniques have been applied in HAR; however, most of them work between domains that share the same feature space; for example, transferring the learn models between users [44] or between sensor wearing positions [32].”
From ContrasGAN: Unsupervised domain adaptation in Human Activity Recognition via adversarial and contrastive learning (2021) · cites this paper as [32]
Experimental — accuracy not yet measuredIn the citing paper
“To address this issue, a representation learning method using triplet loss is proposed to directly learn discriminative and compact face embedding [99].”
From Empowering things with intelligence: a survey of the progress, challenges, and opportunities in artificial intelligence of things (2021) · cites this paper as [99]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Going deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562 · cited 8×
“We would like to thank Johannes Steffens for his discussions and great insights on face recognition and Christian Szegedy for providing new network architectures like [16] and discussing network design choices.”
MethodResultInfluentialLearning representations by back-propagating errors
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986) · Nature · doi:10.1038/323533a0 · Cited by 31,892 · cited 2×
“In all our experiments we train the CNN using Stochastic Gradient Descent (SGD) with standard backprop [8, 11] and AdaGrad [5].”
MethodVisualizing and understanding convolutional networks
Zeiler, M. D., & Fergus, R. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10590-1_53 · Cited by 15,839 · cited 5×
“This table show the structure of our Zeiler&Fergus [22] based model with 1×1 convolutions inspired by [9].”
BackgroundMethodInfluentialBackpropagation Applied to Handwritten Zip Code Recognition
LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989) · Neural Computation · doi:10.1162/neco.1989.1.4.541 · Cited by 11,891 · cited 2×
“In all our experiments we train the CNN using Stochastic Gradient Descent (SGD) with standard backprop [8, 11] and AdaGrad [5].”
Method(2010) · Cited by 8,261
DeepFace: Closing the Gap to Human-Level Performance in Face Verification
Taigman, Y., Yang, M., Ranzato, M., & Wolf, L. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.220 · Cited by 6,542 · cited 9×
“[17], for example, performs a complex 3D alignment.”
BackgroundMethodInfluentialBengio, Y., Louradour, J., Collobert, R., & Weston, J. (2009) · Proceedings of the 26th Annual International Conference on Machine Learning · doi:10.1145/1553374.1553380 · Cited by 5,366 · cited 1×
“Choosing which triplets to use turns out to be very important for achieving good performance and, inspired by curriculum learning [1], we present a novel online negative exemplar mining strategy which ensures consistently increasing difficulty of triplets as the network trains.”
BackgroundLarge Scale Distributed Deep Networks
Dean, J. B., Corrado, G. S., Monga, R., Chen, K., Devin, M., Mao, M., Ranzato, M., Senior, A., Tucker, P. A., Yang, K., et al. (2012) · neural information processing systems · Cited by 4,489
Distance Metric Learning for Large Margin Nearest Neighbor Classification
Weinberger, K. Q., & Saul, L. K. (2009) · Journal of Machine Learning Research · doi:10.5555/1577069.1577078 · Cited by 4,044 · cited 4×
“D embedding using a tripletbased loss function based on LMNN [19].”
MethodInfluentialDeep Learning Face Representation by Joint Identification-Verification
Sun, Y., Wang, X., & Tang, X. (2014) · arXiv · doi:10.48550/arxiv.1406.4773 · Cited by 1,906 · cited 4×
“the one using pairs of positives and negatives, as used in [14] Eq.”
BackgroundMethodInfluential
Research works printed but not linked: 2
- huang, 2007, Labeled faces in the wild: A database for studying face recognition in unconstrained environments, Technical Report, 7
- lin, 2013, Network in network, CaRR abs/1312 4400
Cited by
Items 901–1,000 of 1,094
Face Recognition using FaceNet (Survey, Performance Test, and Comparison)
Iván, W., Setiadi, D. R. I. M., Rachmawanto, E. H., Santoso, H. A., & Sari, C. A. (2019) · International Conference on Informatics and Computing (ICIC) · doi:10.1109/icic47613.2019.8985786 · Cited by 103
“For example, Principal Component Analysis (PCA) [4], Linear Discriminant Analysis (LDA) [5], and Deep Learning [6] [7] [8].”
BackgroundMethodKnowledge-Preserving Incremental Social Event Detection via Heterogeneous GNNs
Cao, Y., Peng, H. M., Wu, J. X., Dou, Y., Li, J., & Yu, P. S. (2021) · International World Wide Web Conference (WWW) · doi:10.1145/3442381.3449834 · Cited by 132
“The triplets are constructed in an online manner, as inspired by computer vision studies [15, 32], to facilitate incremental learning.”
MethodAddressing Imbalance in Multi-Label Classification Using Weighted Cross Entropy Loss Function
Rezaei-Dastjerdehei, M., Mijani, A. M., & Fatemizadeh, E. (2020) · National and International Iranian Conference on Biomedical Engineering (ICBME) · doi:10.1109/icbme51989.2020.9319440 · Cited by 140
“3) In this method training a network, deep metric learning is not only based on accuracy, and other measures such as sensitivity and specificity are applied [24-27].”
MethodUniversal representations:The missing link between faces, text, planktons, and cat breedsPreprint
Bilen, H., & Vedaldi, A. (2017) · Oxford University Research Archive (ORA) (University of Oxford) · doi:10.48550/arxiv.1701.07275 · Cited by 153
“…in a human works well for an enormous variety of different problems, different neural networks are required in order to recognize faces [20, 42, 47, 51], classify [27, 17], detect [44, 35] or segment [9] high-level object categories, read text [15, 23], recognize bird, flower, dog, or cat…”
BackgroundRehman, Y. A. U., Po, L.-M., & Liu, M. (2018) · Expert Systems with Applications · doi:10.1016/j.eswa.2018.05.004 · Cited by 101
Temporal Convolutional Neural Network for the Classification of Satellite Image Time Series
Pelletier, C., Webb, G. I., & Petitjean, F. (2019) · Remote Sensing · doi:10.3390/rs11050523 · Cited by 827
Learning Speaker Representations with Mutual Information
Ravanelli, M., & Bengio, Y. (2019) · Interspeech 2019 · doi:10.21437/interspeech.2019-2380 · Cited by 96
Deep Self-Evolution Clustering
Chang, J., Meng, G., Wang, L., Xiang, S., & Pan, C. (2018) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2018.2889949 · Cited by 104
Melinte, O., & Vlădăreanu, L. (2020) · Sensors · doi:10.3390/s20082393 · Cited by 115
“63% on the widely used labeled faces in the Wild dataset [33].”
BackgroundGroupFace: Learning Latent Groups and Constructing Group-Based Representations for Face Recognition
Kim, Y., Park, W., Roh, M.-C., & Shin, J. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00566 · Cited by 122
“FaceNet [26] proposed triplet loss to maximize the distance between an anchor and its negative sample, and to minimize the distance between the same anchor and its positive sample.”
BackgroundTrust in 5G Open RANs through Machine Learning: RF Fingerprinting on the POWDER PAWR Platform
Reus-Muns, G., Jaisinghani, D., Sankhe, K., & Chowdhury, K. R. (2020) · Globecom · doi:10.1109/globecom42002.2020.9348261 · Cited by 98
“While this work generates the triplets randomly during training time, may other methods are proposed [10], which we will explore in future work.”
BackgroundMethodCycAs: Self-supervised Cycle Association for Learning Re-identifiable Descriptions
Wang, Z., Zhang, J., Zheng, L., Liu, Y., Sun, Y., Li, Y., & Wang, S. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58621-8_5 · Cited by 115
Walk and Learn: Facial Attribute Representation Learning from Egocentric Video and Contextual Data
Wang, J., Cheng, Y., & Feris, R. (2016) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2016.252 · Cited by 115
Improving Cross-database Face Presentation Attack Detection via Adversarial Domain Adaptation
Wang, G., Han, H., Shan, S., & Chen, X. (2019) · International Conference on Biometrics · doi:10.1109/icb45273.2019.8987254 · Cited by 100
Learning to Cluster Faces via Confidence and Connectivity Estimation
Yang, L., Chen, D., Zhan, X., Zhao, R., Loy, C. C., & Lin, D. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01338 · Cited by 122
Deep Hashing for Secure Multimodal Biometrics
Talreja, V., Valenti, M. C., & Nasrabadi, N. M. (2020) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2020.3033189 · Cited by 109
Preventing DeepFake Attacks on Speaker Authentication by Dynamic Lip Movement Analysis
Yang, C.-Z., Ma, J., Wang, S., & Liew, A. W.-C. (2021) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2020.3045937 · Cited by 95
“Very high false acceptance rates were obtained (85% by VGG [47] and 95% by FaceNet [48]) for DeepFake videos.”
BackgroundMehta, N. K., Prasad, S. S., Saurav, S., Saini, R., & Singh, S. (2022) · Applied Intelligence · doi:10.1007/s10489-022-03200-4 · Cited by 97
“As a next step, we use a robust real-time Dlib face detector to crop face regions from each selected frame.”
MethodUse All The Labels: A Hierarchical Multi-Label Contrastive Learning Framework
Zhang, S., Xu, R., Xiong, C., & Ramaiah, C. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01616 · Cited by 157
“In recent years, several unsupervised [5, 13, 16] and supervised [17, 18, 30, 42] metric learning frameworks have been proposed.”
BackgroundMethodMultiple Feature Integration for Classification of Thoracic Disease in Chest Radiography
Ho, T. K., & Gwak, J. (2019) · Applied Sciences · doi:10.3390/app9194130 · Cited by 96
“In the era of computer vision, deep learning-based methods have gained state-of-the-art performances in a variety of domains, including image classification [7,8], scene recognition [9], face recognition [10], and object detection [11].”
BackgroundFace Recognition: Primates in the Wild
Deb, D., Wiper, S., Gong, S., Shi, Y., Tymoszek, C., Fletcher, A., & Jain, A. K. (2018) · 2018 IEEE 9th International Conference on Biometrics Theory, Applications and Systems (BTAS) · doi:10.1109/btas.2018.8698538 · Cited by 87
Deep Metric Learning Beyond Binary Supervision
Kim, S., Seo, M., Laptev, I., Cho, M., & Kwak, S. (2019) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2019.00239 · Cited by 121
“Likewise, understanding similarity between images has played essential roles in many areas of computer vision including image retrieval [19, 43, 44, 50], face identification [12, 39, 46], place recognition [4], pose estimation [45], person reidentification [10, 40], video object tracking [42, 47], local feature descri…”
BackgroundMethodInfluentialGhostVLAD for Set-Based Face Recognition
Zhong, Y., Arandjelović, R., & Zisserman, A. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-20890-5_3 · Cited by 114
Face Attribute Detection with MobileNetV2 and NasNet-Mobile
Saxen, F., Werner, P., Handrich, S., Othman, E., Dinges, L., & Al-Hamadi, A. K. (2019) · International symposium on image and signal processing and analysis/ISPA ... · doi:10.1109/ispa.2019.8868585 · Cited by 104
MMAct: A Large-Scale Dataset for Cross Modal Human Action Understanding
Kong, Q., Wu, Z., Deng, Z., Klinkigt, M., Tong, B., & Murakami, T. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00875 · Cited by 151
Rethinking Triplet Loss for Domain Adaptation
Deng, W., Zheng, L., Sun, Y., & Jiao, J. (2020) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2020.2968484 · Cited by 108
“To preserve the two semantic relations, we naturally implement the constraint by the triplet loss [41].”
BackgroundMethodMulti-Modal Recurrent Attention Networks for Facial Expression Recognition
Lee, J., Kim, S., Kim, S., & Sohn, K. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.2996086 · Cited by 108
“Recently, deep CNN-based methods have substantially outperformed previous handcrafted approaches in various applications such as face recognition [30], facial point detection [31], and face verification [32] with more discriminative features.”
MethodDeep learning based classification of facial dermatological disorders
Goceri, E. (2021) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2020.104118 · Cited by 95
Unsupervised deep clustering via adaptive GMM modeling and optimization
Wang, J., & Jiang, J. (2021) · Neurocomputing · doi:10.1016/j.neucom.2020.12.082 · Cited by 100
Deep graph similarity learning: a survey
Ma, G., Ahmed, N. K., Willke, T. L., & Yu, P. S. (2021) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-020-00733-5 · Cited by 95
“Another widely used loss function for Siamese network is the triplet loss [104].”
BackgroundDual Reweighting Domain Generalization for Face Presentation Attack Detection
Liu, S., Zhang, K.-Y., Yao, T., Sheng, K., Ding, S., Tai, Y. Y., Li, J., Xie, Y., & Ma, L. (2021) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2021/120 · Cited by 117
Deep Meta-Learning: Learning to Learn in the Concept Space
Zhou, F., Wu, B., & Li, Z. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1802.03596 · Cited by 92
“Other methods use metric learning to ease pairwise comparison between examples (Fink, 2005; Koch, 2015; Guillaumin et al., 2009; Schroff et al., 2015).”
MethodDeep Learning for Deepfakes Creation and Detection.Preprint
Nguyen, T. T., Nguyen, C. M., Nguyen, D. T., Nguyen, T. T., & Nahavandi, S. (2019) · arXiv (Cornell University) · Cited by 125
Improving Periocular Recognition by Explicit Attention to Critical Regions in Deep Neural Network
Zhao, Z., & Kumar, A. (2018) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2018.2833018 · Cited by 100
Look, Listen and Learn — A Multimodal LSTM for Speaker Identification
Ren, J., Hu, Y., Tai, Y.-W., Wang, C., Xu, L., Sun, W., & Yan, Q. (2016) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v30i1.10471 · Cited by 91
Unsupervised and Semi-Supervised Domain Adaptation for Action Recognition from Drones
Choi, J., Sharma, G., Chandraker, M. K., & Huang, J. (2020) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv45572.2020.9093511 · Cited by 112
Fashion Outfit Complementary Item Retrieval
Lin, Y.-L., Tran, S., & Davis, L. S. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00337 · Cited by 99
Fine-Grained Spatial Alignment Model for Person Re-Identification With Focal Triplet Loss
Zhou, Q., Zhong, B., Lan, X., Sun, G., Zhang, Y., Zhang, B., & Ji, R. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.3004267 · Cited by 92
“The original triplet loss is proposed in [56] for the face recognition community and then introduced into the person re-ID.”
BackgroundTowards Age-Invariant Face Recognition
Zhao, J., Yan, S., & Feng, J. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.3011426 · Cited by 112
IdentityDP: Differential private identification protection for face images
Wen, Y., Liu, B., Ding, M., Xie, R., & Song, L. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.06.039 · Cited by 116
Bui, T., Ribeiro, L., Ponti, M. A., & Collomosse, J. P. (2018) · Computers & Graphics · doi:10.1016/j.cag.2017.12.006 · Cited by 108
“Triplet CNNs have recently 26 been explored for face identification [28], tracking [29], photo27 graphic visual search in[27, 30] and for sketched queries in or28 der to refine search within a single object class (e.”
Transferring Knowledge Fragments for Learning Distance Metric from a Heterogeneous Domain
Luo, Y., Wen, Y., Liu, T., & Tao, D. (2018) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2018.2824309 · Cited by 100
A CNN-Based Framework for Comparison of Contactless to Contact-Based Fingerprints
Lin, C., & Kumar, A. (2018) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2018.2854765 · Cited by 104
Solving Mixed-Modal Jigsaw Puzzle for Fine-Grained Sketch-Based Image Retrieval
Pang, K., Yang, Y., Hospedales, T. M., Xiang, T., & Song, Y.-Z. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01036 · Cited by 107
“Pre-training + Fine-tuning Many deep CNN based computer vision models assume that a rich universal representation has been captured in ImageNet pre-trained CNN [31, 7, 35, 26], which can then be fined-tuned with task-specific data using various strategies [13, 30, 25, 21, 8].”
MethodMutual Contrastive Learning for Visual Representation Learning
Yang, C., An, Z., Cai, L., & Xu, Y. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i3.20211 · Cited by 115
Automatic soiling and partial shading assessment on PV modules through RGB images analysis
Cavieres, R., Barraza, R., Estay, D., Bilbao, J. I., & Valdivia, P. (2021) · Applied Energy · doi:10.1016/j.apenergy.2021.117964 · Cited by 112
An, X., Deng, J., Guo, J., Feng, Z., Zhu, X., Yang, J., & Liu, T. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.00401 · Cited by 152
“[13, 30], and the design of margin-based and mining-based loss functions [8, 17, 25, 30, 35, 36, 38, 41].”
BackgroundDLFace: Deep local descriptor for cross-modality face recognition
Peng, C., Wang, N., Li, J., & Gao, X. (2019) · Pattern Recognition · doi:10.1016/j.patcog.2019.01.041 · Cited by 95
One-shot Face Recognition by Promoting Underrepresented Classes
Guo, Y., & Zhang, L. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1707.05574 · Cited by 90
“The closest runner-up to our CCS-face is FaceNet in [18], which is trained with > 100M private images for 8M persons, while our model is reproducible.”
MethodDeep Neural Networks with Relativity Learning for facial expression recognition
Guo, Y., Tao, D., Yu, J., Xiong, H., Li, Y., & Tao, D. (2016) · 2016 IEEE International Conference on Multimedia & Expo Workshops (ICMEW) · doi:10.1109/icmew.2016.7574736 · Cited by 93
“Given a network and treating it as an end-to-end system, triplet loss [13] can be directly applied to the practical problem of expression recognition; that is, triplet loss training aims to learn an optimal feature representation that maximizes the relative distance.”
BackgroundMulti-Person Articulated Tracking With Spatial and Temporal Embeddings
Jin, S., Liu, W., Ouyang, W., & Qian, C. (2019) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2019.00581 · Cited by 105
Text2Scene: Generating Compositional Scenes From Textual Descriptions
Tan, F., Feng, S., & Ordonez, V. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00687 · Cited by 93
Gaussian Affinity for Max-Margin Class Imbalanced Learning
Hayat, M., Khan, S. H., Zamir, S. W., Shen, J., & Shao, L. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00657 · Cited by 123
Incomplete Descriptor Mining With Elastic Loss for Person Re-Identification
Tan, H., Liu, X., Bian, Y., Wang, H., & Yin, B. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3061412 · Cited by 109
Variational Prototype Learning for Deep Face Recognition
Deng, J., Guo, J., Yang, J., Lattas, A., & Zafeiriou, S. P. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01173 · Cited by 126
“The (N+1)-tuplet is defined as follows: When N = 2, the (2+1)-tuplet loss highly resembles the triplet loss [35] as there is only one negative sample for each positive pair.”
BackgroundMethodInfluentialDE-GAN: Domain Embedded GAN for High Quality Face Image Inpainting
Zhang, X., Wang, X., Shi, C., Yan, Z., Li, X., Kong, B., Lyu, S., Zhu, B. B., Lv, J., Yin, Y., et al. (2021) · Pattern Recognition · doi:10.1016/j.patcog.2021.108415 · Cited by 105
Shad, H. S., Rizvee, M. M., Roza, N. T., Hoq, S. M. A., Monirujjaman Khan, M., Singh, A., Zaguia, A., & Bourouis, S. (2021) · Computational Intelligence and Neuroscience · doi:10.1155/2021/3111676 · Cited by 93
RetractedQu, F., Dang, N., Furht, B., & Nojoumian, M. (2024) · Journal Of Big Data · doi:10.1186/s40537-024-00890-0 · Cited by 120
“M e t h o d o l o g y R e l e v a n c e [ 65 ] U t i l i z e s f e a t u r e v e c t o r s i n F a c e N e t f o r e ffi c i e n t f a c e r e c o g n i t i o n , c l u s t e r i n g , a n d v e r i fi c a t i o n t a s k s .”
Failures of Gradient-Based Deep Learning
Shalev‐Shwartz, S., Shamir, O., & Shammah, S. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.07950 · Cited by 62
“Introduction The success stories of deep learning form an ever lengthening list of practical breakthroughs and state-of-theart performances, ranging the fields of computer vision (Krizhevsky et al., 2012; He et al., 2016; Schroff et al., 2015; Taigman et al., 2014), audio and natural language processing and generation…”
BackgroundLocal binary pattern network: A deep learning approach for face recognition
Xi, M., Chen, L., Polajnar, D., & Tong, W. (2016) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2016.7532955 · Cited by 102
“FaceNet [14]) which achieve superior results; however, they used a less restricted setting: external data are allowed for training.”
Zhu, Q., Zhong, Y., Zhang, L., & Li, D. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2833293 · Cited by 99
“It has been successfully applied to objection detection [19], facial recognition [20], hyperspectral pixel classification and reconstruction [21]–[23], and video analysis [24].”
MethodFew-Shot Deep Adversarial Learning for Video-Based Person Re-Identification
Wu, L. Y., Wang, Y., Yin, H., Wang, M., & Shao, L. (2019) · IEEE Transactions on Image Processing · doi:10.1109/tip.2019.2940684 · Cited by 101
Deep Temporal-Spatial Feature Learning for Motor Imagery-Based Brain–Computer Interfaces
Chen, J., Yu, Z., Gu, Z., & Li, Y. (2020) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2020.3023417 · Cited by 95
“In this paper, we perform the online selection of mini-batch triplets in the training process [32].”
MethodInfluentialMIANet: Aggregating Unbiased Instance and General Information for Few-Shot Semantic Segmentation
Yang, Y., Chen, Q., Feng, Y., & Huang, T. S. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00689 · Cited by 122
“Next, a well-designed triplet loss [25] is applied to achieve the alignment between the semantic prototype and the visual features.”
MethodHistogram-based local descriptors for facial expression recognition (FER): A comprehensive study
Turan, C., & Lam, K. (2018) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2018.05.024 · Cited by 98
Transform and Tell: Entity-Aware News Image Captioning
Tran, A., Mathews, A., & Xie, L. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.01305 · Cited by 87
“A vector representation of each face is obtained by passing the bounding boxes to FaceNet [40], which was pre-trained on the VGGFace2 dataset [5].”
MethodA Self-boosting Framework for Automated Radiographic Report Generation
Wang, Z., Zhou, L., Wang, L., & Li, X. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00246 · Cited by 83
Nazir, A., Cheema, M. N., Sheng, B., Li, P., Li, H., Xue, G., Qin, J., Kim, J., & Feng, D. D. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3136619 · Cited by 99
Weakly Supervised Temporal Action Localization via Representative Snippet Knowledge Propagation
Huang, L., Wang, L., & Li, H. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.00327 · Cited by 115
“For example, W-TALC [36], 3C-Net [32], RPN [7] and A2CL-PT [29] employed the center loss [44], clustering loss [48], triplet loss [40], etc.”
BackgroundPart-Based Deep Hashing for Large-Scale Person Re-Identification
Zhu, F., Kong, X., Zheng, L., Fu, H., & Tian, Q. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2695101 · Cited by 96
The Do’s and Don’ts for CNN-Based Face Verification
Bansal, A., Castillo, C. D., Ranjan, R., & Chellappa, R. (2017) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2017.299 · Cited by 113
Attention-based Pyramid Aggregation Network for Visual Place Recognition
Zhu, Y., Wang, J., Xie, L., & Zheng, L. (2018) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3240508.3240525 · Cited by 95
Centralized Large Margin Cosine Loss for Open-Set Deep Palmprint Recognition
Zhong, D., & Zhu, J. (2019) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2019.2904283 · Cited by 102
“Multiple verification losses, such as contrastive loss [43] and triplet loss [5], are proposed to enforce intra-class compactness and inter-class dispersion in Euclidean space.”
BackgroundDetecting Deepfakes with Metric Learning
Kumar, A., Bhavsar, A., & Verma, R. (2020) · 2020 8th International Workshop on Biometrics and Forensics (IWBF) · doi:10.1109/iwbf49977.2020.9107962 · Cited by 90
“Then, we used FaceNet [17] to generate face embeddings of a dimension of 512.”
MethodMulti-Model Long Short-Term Memory Network for Gait Recognition Using Window-Based Data Segment
Tran, L., Hoang, T., Nguyen, T., Kim, H., & Choi, D. (2021) · IEEE Access · doi:10.1109/access.2021.3056880 · Cited by 90
“• We further extend the proposed network to build the verification model with triplet loss [25] and One-class Support Vector Machine (OCSVM) [26].”
MethodAn ultrafast bipolar flash memory for self-activated in-memory computing
Huang, X., Liu, C., Tang, Z., Zeng, S., Wang, S., & Zhou, P. (2023) · Nature Nanotechnology · doi:10.1038/s41565-023-01339-w · Cited by 107
Toward Unified Token Learning for Vision-Language Tracking
Zheng, Y., Zhong, B., Liang, Q., Li, G., Ji, R., & Li, X. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3301933 · Cited by 165
“For instance, they often use various loss functions (such as triplet loss [10], focal loss [11], GIoU loss [12], and diversity loss [13]) to supervise the proposed modules or tasks with specific functions.”
BackgroundMS-Celeb-1M: Challenge of Recognizing One Million Celebrities in the Real World
Guo, Y., Zhang, L., Hu, Y., He, X., & Gao, J. (2016) · Electronic Imaging · doi:10.2352/issn.2470-1173.2016.11.imawm-463 · Cited by 89
Discriminative Transfer Feature and Label Consistency for Cross-Domain Image Classification
Li, S., Liu, C. H., Su, L., Xie, B., Ding, Z., Chen, C. L. P., & Wu, D. O. (2020) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2019.2958152 · Cited by 94
“Inspired by the triplet loss [27], [45], which minimizes the distance of the within-class sample pairs while maximizing the distance of the between-class sample pairs in feature embedding space, we only focus on the hardest sample pairs to optimize as in Fig.”
BackgroundMethodRecent development in face recognition
Jayaraman, U., Gupta, P., Gupta, S., Arora, G., & Tiwari, K. (2020) · Neurocomputing · doi:10.1016/j.neucom.2019.08.110 · Cited by 100
Serna, I., Morales, A., Fiérrez, J., & Obradovich, N. (2022) · Artificial Intelligence · doi:10.1016/j.artint.2022.103682 · Cited by 118
A deep learning model for detection and tracking in high-throughput images of organoid
Bian, X., Li, G., Wang, C., Liu, W., Lin, X., Chen, Z., Cheung, M., & Luo, X. (2021) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2021.104490 · Cited by 88
Personalized and Invertible Face De-identification by Disentangled Identity Information Manipulation
Cao, J., Liu, B., Wen, Y., Xie, R., & Song, L. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00332 · Cited by 112
Mei, S., Jiang, R., Ma, M., & Song, C. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3233726 · Cited by 103
“In the object detection [8] and classification [9], [10]”
Forensic Face Photo-Sketch Recognition Using a Deep Learning-Based Architecture
Galea, C., & Farrugia, R. A. (2017) · IEEE Signal Processing Letters · doi:10.1109/lsp.2017.2749266 · Cited by 89
“Consequently, such networks must be trained using not only a vast number of classes, but also numerous images for each category [8], [9].”
BackgroundLarge Age-Gap face verification by feature injection in deep networks
Bianco, S. (2017) · Pattern Recognition Letters · doi:10.1016/j.patrec.2017.03.006 · Cited by 95
Scene Classification Based on a Deep Random-Scale Stretched Convolutional Neural Network
Liu, Y., Zhong, Y., Feng, F., Zhu, Q., & Qin, Q. (2018) · Remote Sensing · doi:10.3390/rs10030444 · Cited by 89
Sitcom-star-based clothing retrieval for video advertising: a deep learning framework
Zhang, H., Ji, Y., Huang, W., & Liu, L. (2018) · Neural Computing and Applications · doi:10.1007/s00521-018-3579-x · Cited by 92
“[40] introduced a triplet loss in a novel face verification model, FaceNet.”
BackgroundRethinking the Detection of Child Sexual Abuse Imagery on the Internet
Bursztein, E., Clarke, E., DeLaune, M., Elifff, D. M., Hsu, N., Olson, L., Shehan, J., Thakur, M., Thomas, K., & Bright, T. (2019) · The World Wide Web Conference · doi:10.1145/3308558.3313482 · Cited by 80
User Authentication Based on Mouse Dynamics Using Deep Neural Networks: A Comprehensive Study
Chong, P., Elovici, Y., & Binder, A. (2019) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2019.2930429 · Cited by 96
“In [1], the researchers measured facial similarity by learning a mapping from facial images to a compact”
BackgroundContrastive-center loss for deep neural networks
Qi, C., & Su, F. (2017) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2017.8296803 · Cited by 109
ShuffleFaceNet: A Lightweight Face Architecture for Efficient and Highly-Accurate Face Recognition
Martínez-Díaz, Y., Luévano, L. S., Méndez-Vázquez, H., Nicolás-Díaz, M., Chang, L., & González-Mendoza, M. (2019) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2019.00333 · Cited by 111
Face Recognition and Classification Using GoogleNET Architecture
Anand, R., Shanthi, T. S., Nithish, M. S., et al. (2019) · Advances in intelligent systems and computing · doi:10.1007/978-981-15-0035-0_20 · Cited by 88
He, J., Wang, Y., & Liu, H. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.3009284 · Cited by 97
Ranasinghe, K., Naseer, M., Hayat, M., Khan, S. H., & Khan, F. S. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01211 · Cited by 120
“Contrastive and triplet loss functions [16, 45] are dependent on carefully designed negative mining procedures, which are both time-consuming and performance-sensitive.”
Sun, W., Dai, G., Zhang, X., He, X., & Chen, X. (2021) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2021.3130403 · Cited by 93
Praveen, S. P., Nakka, R., Anuradha, C., Thatha, V. N., Vellela, S. S., & Sirisha, U. (2023) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2023.01406128 · Cited by 88
“…into classifying agricultural diseases uses a wide range of approaches, including “genetic algorithms” [6], “sup-port vector machines” [7],“K-means clustering” [8], “ensemble learning” [9], “Bayesian classification” [10], “radial basis functions” [11], & “filter segmentation” techniques [12].”
Automated Radiology Report Generation: A Review of Recent Advances
Sloan, P., Clatworthy, P. L., Simpson, E., & Mirmehdi, M. (2024) · IEEE Reviews in Biomedical Engineering · doi:10.1109/rbme.2024.3408456 · Cited by 169
“Although not as commonly applied as contrastive loss, triplet loss [80] has been used in ARRG for training the model encoder, and as an auxiliary task to improve a model’s performance.”
Few-shot learning for short text classification
Yan, L., Zheng, Y., & Cao, J. (2018) · Multimedia Tools and Applications · doi:10.1007/s11042-018-5772-4 · Cited by 116
“[28] propose a system FaceNet, which learns a mapping between image space and Euclidean space to directly measure the similarity.”
BackgroundAn End-to-End System for Unconstrained Face Verification with Deep Convolutional Neural Networks
Chen, J.-C., Ranjan, R., Kumar, A., Chen, C., Patel, V. M., & Chellappa, R. (2015) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2015.55 · Cited by 99
“Keywords deep learning · face detection/association · fiducial detection · face verification · metric learning ∗ First five authors contributed equally Jun-Cheng Chen A.V. Williams 4455, University of Maryland, College Park, MD 20740, USA E-mail: pullpull@cs.umd.edu”
BackgroundMethod
Metadata
- Authors
- Schroff, F., Kalenichenko, D., & Philbin, J.
- Affiliations
- Google (United States) (US)
- Venue
- 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2015-06
- Pages
- 815-823
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W2096733369
- Semantic Scholar
- 5aa26299435bdf7db874ef1640a6c3b5a4a2c394
- Cited by
- 15,267
- References
- 22
- Citations by year
- 2026: 846 · 2025: 1556 · 2024: 1590 · 2023: 1823 · 2022: 1872 · 2021: 2113 · 2020: 1902 · 2019: 1574 · 2018: 1062 · 2017: 593 · 2016: 281 · 2015: 51
- FWCI
- 271.65
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (publishedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face and Expression Recognition
- Life Sciences › Neuroscience › Cognitive Neuroscience › Face Recognition and Perception
- Keywords
- FaceNet, face recognition, face verification, face clustering, embedding space, triplet loss, deep convolutional neural networks, face similarity, Labeled Faces in the Wild
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.