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 401–500 of 1,094
Deep Transfer Learning for Person Re-identification
Geng, M., Wang, Y., Xiang, T., & Tian, Y. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1611.05244 · Cited by 192
Context-Aware Local Binary Feature Learning for Face Recognition
Duan, Y., Lu, J., Feng, J., & Zhou, J. (2018) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2017.2710183 · Cited by 205
αDiff: cross-version binary code similarity detection with DNN
Liu, B., Huo, W., Zhang, C., Li, W., Li, F., Piao, A., & Zou, W. (2018) · Proceedings of the 33rd ACM/IEEE International Conference on Automated Software Engineering · doi:10.1145/3238147.3238199 · Cited by 192
“Similar to [37, 44, 47], we also generate negative samples in each mini-batch during training, based on the positive samples.”
BackgroundMethodResultInfluentialSpherical Kernel for Efficient Graph Convolution on 3D Point Clouds
Lei, H., Akhtar, N., & Mian, A. (2021) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.2983410 · Cited by 202
Composite Backdoor Attack for Deep Neural Network by Mixing Existing Benign Features
Lin, J., Xu, L., Liu, Y., & Zhang, X. (2020) · Proceedings of the 2020 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/3372297.3423362 · Cited by 196
“Another more general task, called face verification [42, 47, 49], is to use the model as an encoder that encodes a face image to a feature vector so that images belonging to a same person have similar vector values.”
BackgroundDeep learning‐based methods for individual recognition in small birds
Ferreira, A. C., Silva, L. R., Renna, F., Brandl, H. B., Renoult, J. P., Farine, D. R., Covas, R., & Doutrelant, C. (2020) · Methods in Ecology and Evolution · doi:10.1111/2041-210x.13436 · Cited by 203
Deep convolutional learning for Content Based Image Retrieval
Tzelepi, M., & Tefas, A. (2018) · Neurocomputing · doi:10.1016/j.neucom.2017.11.022 · Cited by 201
Deep Adversarial Metric Learning
Duan, Y., Zheng, W., Lin, X., Lu, J., & Zhou, J. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00294 · Cited by 193
“[28] selected “semi-hard” negative samples to train FaceNet with triplet loss, which are hard but still farther than the positive-anchor pairs.”
BackgroundMethodA Zero-Shot Framework for Sketch Based Image Retrieval
Yelamarthi, S. K., Reddy, S. K., Mishra, A., & Mittal, A. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-01225-0_19 · Cited by 198
FaceQnet: Quality Assessment for Face Recognition based on Deep Learning
Hernandez-Ortega, J., Galbally, J., Fierrez, J., Haraksim, R., & Beslay, L. (2019) · 2019 International Conference on Biometrics (ICB) · doi:10.1109/icb45273.2019.8987255 · Cited by 191
“After that, we employed the FaceNet pretrained model for feature extraction [16], and we obtained all the mated scores using the Euclidean Distance between the embeddings of the ICAO-compliant gallery images and the rest of the images of the same subject.”
BackgroundMethodInfluentialAffective Image Content Analysis: Two Decades Review and New Perspectives
Zhao, S., Yao, X., Yang, J., Jia, G., Ding, G., Chua, T.-S., Schuller, B. W., & Keutzer, K. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2021.3094362 · Cited by 198
“Based on the triplet loss [125], Yang et al.”
Channel Interaction Networks for Fine-Grained Image Categorization
Gao, Y., Han, X., Wang, X., Huang, W., & Scott, M. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i07.6712 · Cited by 188
Deep Hierarchical Semantic Segmentation
Li, L., Zhou, T., Wang, W., Li, J., & Yang, Y. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00131 · Cited by 184
“Note that this is different from the classic, hierarchy-agnostic triplet loss [66], where the anchor and positive samples are from the same class, while the anchor and negative samples are from different classes, i.”
BackgroundDeep Learning for Understanding Faces: Machines May Be Just as Good, or Better, than Humans
Ranjan, R., Sankaranarayanan, S., Bansal, A., Bodla, N., Chen, J.-C., Patel, V. M., Castillo, C. D., & Chellappa, R. (2018) · IEEE Signal Processing Magazine · doi:10.1109/msp.2017.2764116 · Cited by 197
“72 IEEE SIgnal ProcESSIng MagazInE | January 2018 | Schroff et al. proposed a CNN-based approach for face recognition called FaceNet [62] that directly optimizes the embedding itself, rather than an intermediate bottleneck layer as in other deep-learning approaches.”
BackgroundMethodInfluentialAccurate iris segmentation in non-cooperative environments using fully convolutional networks
Liu, N., Li, H., Zhang, M., Liu, J., Sun, Z., & Tan, T. (2016) · 2016 International Conference on Biometrics (ICB) · doi:10.1109/icb.2016.7550055 · Cited by 187
Improved Deep Hashing With Soft Pairwise Similarity for Multi-Label Image Retrieval
Zhang, Z., Zou, Q., Lin, Y., Chen, L., & Wang, S. (2020) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2019.2929957 · Cited by 195
Wang, W., Cui, Z., Yan, Y., Feng, J., Yan, S., Shu, X., & Sebe, N. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.261 · Cited by 185
A Fast and Accurate System for Face Detection, Identification, and Verification
Ranjan, R., Bansal, A., Zheng, J., Xu, H., Gleason, J., Lu, B., Nanduri, A., Chen, J.-C., Castillo, C., & Chellappa, R. (2019) · IEEE Transactions on Biometrics, Behavior, and Identity Science · doi:10.1109/tbiom.2019.2908436 · Cited by 195
Discriminability Objective for Training Descriptive Captions
Luo, R., Shakhnarovich, G., Cohen, S., & Price, B. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00728 · Cited by 191
Towards Transferable Adversarial Attack Against Deep Face Recognition
Zhong, Y., & Deng, W. (2021) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2020.3036801 · Cited by 173
“For a target black-box deep face model, there are a wide range of options for its training databases [30], [31], [32], [33], training loss functions [1], [2], [3], [4], [5], [6], [7] and network architectures [34], [35], [36], [37].”
BackgroundMethodInfluentialA Survey on Deep Visual Place Recognition
Masone, C., & Caputo, B. (2021) · IEEE Access · doi:10.1109/access.2021.3054937 · Cited by 192
“to overfitting [83] and bad local minima [98].”
BackgroundMultiaccuracy: Black-Box Post-Processing for Fairness in Classification
Kim, M. P., Ghorbani, A., & Zou, J. (2019) · Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3306618.3314287 · Cited by 193
Exploiting Unlabeled Data in CNNs by Self-Supervised Learning to Rank
Liu, X., Weijer, J. V. D., & Bagdanov, A. D. (2019) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2899857 · Cited by 192
“In [50] the authors propose semi-hard pair selection, arguing that selecting hardest pairs can lead to bad local minima.”
BackgroundAnalysis Survey on Deepfake detection and Recognition with Convolutional Neural Networks
Ahmed, S. R., Sonuc, E., Ahmed, M. R., & Duru, A. D. (2022) · 2022 International Congress on Human-Computer Interaction, Optimization and Robotic Applications (HORA) · doi:10.1109/hora55278.2022.9799858 · Cited by 187
Image Search With Text Feedback by Visiolinguistic Attention Learning
Chen, Y., Gong, S., & Bazzani, L. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00307 · Cited by 184
“We adopt semihard mining [54] to select the negative pair x̄n.”
MethodPose-Guided Photorealistic Face Rotation
Hu, Y., Wu, X., Yu, B., He, R., & Sun, Z. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00876 · Cited by 187
“For the first class, metric learning [25] is a common way to achieve pose-invariant embeddings.”
BackgroundUnsupervised Discovery of Object Landmarks as Structural Representations
Zhang, Y., Guo, Y., Jin, Y., Luo, Y., He, Z., & Lee, H. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00285 · Cited by 186
Class Rectification Hard Mining for Imbalanced Deep Learning
Dong, Q., Gong, S., & Zhu, X. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.205 · Cited by 189
“Negative mining has been used for pedestrian detection [14], face recognition [38], image categorisation [35, 46, 9], unsupervised visual representation learning [48].”
MethodDense Human Body Correspondences Using Convolutional Networks
Wei, L., Huang, Q., Ceylan, D., Vouga, E., & Li, H. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.171 · Cited by 185
MIN2Net: End-to-End Multi-Task Learning for Subject-Independent Motor Imagery EEG Classification
Autthasan, P., Chaisaen, R., Sudhawiyangkul, T., Rangpong, P., Kiatthaveephong, S., Dilokthanakul, N., Bhakdisongkhram, G., Phan, H., Guan, C., & Wilaiprasitporn, T. (2022) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2021.3137184 · Cited by 187
“…from particular frequency bands (mu (9–13 Hz) and beta (22–29 Hz ) ) In recent years, the use of Deep Learning (DL) has been shown promising results and proven itself to be a successful set of models in the field of computer vision, speech recognition, and natural language processing [18], [19].”
BackgroundTransVPR: Transformer-Based Place Recognition with Multi-Level Attention Aggregation
Wang, R., Shen, Y., Zuo, W., Zhou, S., & Zheng, N. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.01328 · Cited by 185
Semantic Instance Segmentation via Deep Metric Learning
Fathi, A., Wojna, Z., Rathod, V., Wang, P., Song, H. O., Guadarrama, S., & Murphy, K. P. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.10277 · Cited by 182
“This is similar to other approaches, such as FaceNet [25], which learn how likely two bounding boxes are to belong to the same instance (person), except we learn to predict the similarity of pixels , taking into account their local context.”
MethodGraphIQA: Learning Distortion Graph Representations for Blind Image Quality Assessment
Sun, S., Yu, T., Xu, J., Zhou, W., & Chen, Z. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3152942 · Cited by 185
Level Playing Field for Million Scale Face Recognition
Nech, A., & Kemelmacher-Shlizerman, I. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.363 · Cited by 180
“Optimizing directly for embeddings, without predicting all classes at once, showed to scale better [25].”
InfluentialTowards More Accurate Iris Recognition Using Deeply Learned Spatially Corresponding Features
Zhao, Z., & Kumar, A. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.411 · Cited by 187
Text2Shape: Generating Shapes from Natural Language by Learning Joint Embeddings
Chen, K., Choy, C. B., Savva, M., Chang, A. X., Funkhouser, T., & Savarese, S. (2019) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-20893-6_7 · Cited by 186
SensitiveNets: Learning Agnostic Representations with Application to Face Images
Morales, A., Fierrez, J., Vera-Rodriguez, R., & Tolosana, R. (2021) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.3015420 · Cited by 186
“triplets where the margin between the inter-class and the intra-class distances is limited by a as proposed in [21], [22].”
BackgroundMethodData-Distortion Guided Self-Distillation for Deep Neural Networks
Xu, T.-B., & Liu, C.-L. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33015565 · Cited by 178
“…the common data-distortion techniques (e.g., random mirror/cropping, rotation, multiple scales) only enlarge additional virtual samples from the neighborhood of existing training samples, which makes network see more distorted training samples to improve model generalization (Simard et al. 1996).”
BackgroundUnsupervised Tracklet Person Re-Identification
Li, M., Zhu, X., & Gong, S. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2903058 · Cited by 185
Mis-Classified Vector Guided Softmax Loss for Face Recognition
Wang, X., Zhang, S., Wang, S., Fu, T., Shi, H., & Mei, T. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i07.6906 · Cited by 181
Unsupervised Pretraining Transfers Well Across Languages
Riviere, M., Joulin, A., Mazare, P.-E., & Dupoux, E. (2020) · ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp40776.2020.9054548 · Cited by 186
“for speech and we focus on those based on contrastive learning [7, 14, 15].”
BackgroundSigNet: Convolutional Siamese Network for Writer Independent Offline Signature Verification
Dey, S., Dutta, A., Toledo, J. I., Ghosh, S. K., Lladós, J., & Pal, U. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1707.02131 · Cited by 173
“This framework has been successfully used for dimensionality reduction in weakly supervised metric learning [15] and for face verification in [16].”
BackgroundMethodSupervised Transformer Network for Efficient Face Detection
Chen, D., Hua, G., Wen, F., & Sun, J. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46454-1_8 · Cited by 185
An introduction to deep learning in medical physics: advantages, potential, and challenges
Shen, C., Nguyen, D., Zhou, Z., Jiang, S. B., Dong, B., & Jia, X. (2020) · Physics in Medicine & Biology · doi:10.1088/1361-6560/ab6f51 · Cited by 185
“, 2014), face recognition (Schroff et al., 2015), image classification (Krizhevsky et al.”
BackgroundMomentum contrastive learning for few-shot COVID-19 diagnosis from chest CT images
Chen, X., Yao, L., Zhou, T., Dong, J., & Zhang, Y. (2021) · Pattern Recognition · doi:10.1016/j.patcog.2021.107826 · Cited by 185
Sample4Geo: Hard Negative Sampling For Cross-View Geo-Localisation
Deuser, F., Habel, K., & Oswald, N. (2023) · 2023 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01545 · Cited by 175
“Deep metric learning for cross-view geo-localisation uses Triplet loss by default [43], with various extensions such as soft-margin triplet loss [34] or weighted soft-margin triplet loss [15].”
BackgroundAttention-Aware Deep Reinforcement Learning for Video Face Recognition
Rao, Y., Lu, J., & Zhou, J. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.424 · Cited by 182
“To train the recurrent layer, we adopt the triplet loss [32] with the cosine similarity to minimize the following objective:”
BackgroundMethodInfluentialContinuous Authentication Through Finger Gesture Interaction for Smart Homes Using WiFi
Kong, H., Lu, L., Yu, J., Chen, Y., & Tang, F. (2021) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2020.2994955 · Cited by 184
Text-Independent Speaker Verification Based on Triplet Convolutional Neural Network Embeddings
Zhang, C., Koishida, K., & Hansen, J. H. L. (2018) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2018.2831456 · Cited by 183
“In that study, the triplet loss is employed to optimize the network training [23], where”
BackgroundMethodResultInfluentialRecent advancements in multimodal human–robot interaction
Su, H., Qi, W., Chen, J., Yang, C., Sandoval, J., & Laribi, M. A. (2023) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2023.1084000 · Cited by 181
“Facial recognition: Robots can use computer vision to recognize and identify specific individuals by analyzing their facial features (Schroff et al., 2015).”
BackgroundA Cookbook of Self-Supervised Learning
Balestriero, R., Ibrahim, M., Sobal, V., Morcos, A., Shekhar, S., Goldstein, T., Bordes, F., Bardes, A., Mialon, G., Tian, Y., et al. (2023) · arXiv · doi:10.48550/arxiv.2304.12210 · Cited by 182
“…examples from different classes should be larger than m. Similar to the contrastive loss, the Triplet loss [Weinberger and Saul, 2009, Chechik et al., 2010, Schroff et al., 2015] shares a similar spirit, but is composed of triplets: a query, a positive example, and a negative example (see eq.”
BackgroundInfluentialTowards Metrical Reconstruction of Human Faces
Zielonka, W., Bolkart, T., & Thies, J. (2022) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-19778-9_15 · Cited by 175
Striking the Right Balance With Uncertainty
Khan, S., Hayat, M., Zamir, S. W., Shen, J., & Shao, L. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00019 · Cited by 180
“Note that some of the compared methods used as much as 200M images [43] and an ensemble of 25 models [48] for training.”
MethodFaceTime — Deep learning based face recognition attendance system
Arsenovic, M., Sladojevic, S., Anderla, A., & Stefanovic, D. (2017) · 2017 IEEE 15th International Symposium on Intelligent Systems and Informatics (SISY) · doi:10.1109/sisy.2017.8080587 · Cited by 177
jTrans: jump-aware transformer for binary code similarity detection
Wang, H., Qu, W., Katz, G., Zhu, W., Gao, Z., Qiu, H., Zhuge, J., & Zhang, C. (2022) · Proceedings of the 31st ACM SIGSOFT International Symposium on Software Testing and Analysis · doi:10.1145/3533767.3534367 · Cited by 169
“The objective function for our fine-tuning process in trained using contrastive learning [25, 53] and is performed as follows:”
MethodGiT: Graph Interactive Transformer for Vehicle Re-Identification
Shen, F., Xie, Y., Zhu, J., Zhu, X., & Zeng, H. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3238642 · Cited by 178
A Boundary Tilting Persepective on the Phenomenon of Adversarial Examples
Tanay, T., & Griffin, L. D. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1608.07690 · Cited by 121
“As a consequence, deep networks that are able to outperform humans are now being produced: for instance on the challenging imageNet dataset (He et al., 2015b), or on face recognition (Schroff et al., 2015).”
BackgroundDas, A., Dantcheva, A., & Bremond, F. (2019) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-11009-3_35 · Cited by 167
“The Python library of Facenet1 is used to calculate facial embedding of face images and developing the proposed MTCNN.”
MethodInfluentialLee, J., Kim, E., Lee, Y., Kim, D., Chang, J., & Choo, J. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00584 · Cited by 171
Partially View-aligned Representation Learning with Noise-robust Contrastive Loss
Yang, M., Li, Y., Huang, Z., Liu, Z., Hu, P., & Peng, X. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00119 · Cited by 171
Facial expression recognition based on deep learning
Ge, H., Zhu, Z., Dai, Y., Wang, B., & Wu, X. (2022) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2022.106621 · Cited by 178
Visual Object Tracking with Discriminative Filters and Siamese Networks: A Survey and Outlook
Javed, S., Danelljan, M., Khan, F. S., Khan, M. H., Felsberg, M., & Matas, J. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3212594 · Cited by 175
Cross-Modal and Hierarchical Modeling of Video and Text
Zhang, B., Hu, H., & Sha, F. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-01261-8_23 · Cited by 173
“These losses are margin-based losses [29] where α and β are positive numbers as the margins to separate matched pairs from unmatched ones.”
BackgroundAdaptiveFace: Adaptive Margin and Sampling for Face Recognition
Liu, H., Zhu, X., Lei, Z., & Li, S. Z. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.01222 · Cited by 170
Robust LSTM-Autoencoders for Face De-Occlusion in the Wild
Zhao, F., Feng, J., Zhao, J., Yang, W., & Yan, S. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2771408 · Cited by 175
Towards Universal Representation Learning for Deep Face Recognition
Shi, Y., Yu, X., Sohn, K., Chandraker, M., & Jain, A. K. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00685 · Cited by 169
Visual identification of individual Holstein-Friesian cattle via deep metric learning
Andrew, W., Gao, J., Mullan, S., Campbell, N., Dowsey, A. W., & Burghardt, T. (2021) · Computers and Electronics in Agriculture · doi:10.1016/j.compag.2021.106133 · Cited by 172
Unsupervised Feature Learning via Non-Parametric Instance-level Discrimination
Wu, Z., Xiong, Y., Yu, S. X., & Lin, D. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1805.01978 · Cited by 163
“An important technical point on metric learning for face recognition is normalization [35, 22, 43], which we also utilize in this work.”
BackgroundMethodAttribute Attention for Semantic Disambiguation in Zero-Shot Learning
Liu, Y., Guo, J., Cai, D., & He, X. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00680 · Cited by 175
Multimodal Measurement of Depression Using Deep Learning Models
Yang, L., Jiang, D., Xia, X., Pei, E., Oveneke, M. C., & Sahli, H. (2017) · Proceedings of the 7th Annual Workshop on Audio/Visual Emotion Challenge · doi:10.1145/3133944.3133948 · Cited by 167
“Deep Convolutional Neural Network (DCNN) has been increasingly used and has demonstrated impressive performances on different tasks such as object recognition [24], object detection [8], face verification [21], emotion recognition [17] [6] [27] and depression classification [16] [29].”
MethodAttributes as Operators: Factorizing Unseen Attribute-Object Compositions
Nagarajan, T., & Grauman, K. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-01246-5_11 · Cited by 167
Embedding Adversarial Learning for Vehicle Re-Identification
Lou, Y., Bai, Y., Liu, J., Wang, S., & Duan, L.-Y. (2019) · IEEE Transactions on Image Processing · doi:10.1109/tip.2019.2902112 · Cited by 175
“For example, a desired embedding model should be discriminative enough to scrupulously observe the subtle differences between the visually similar vehicles with different IDs, and increasing the number of hard negatives that break the distance constraints for triplet constitution has been proven to be pivotal for trai…”
BackgroundMethodSub-Graph Contrast for Scalable Self-Supervised Graph Representation Learning
Jiao, Y., Xiong, Y., Zhang, J., Zhang, Y., Zhang, T., & Zhu, Y. (2020) · Proceedings · doi:10.1109/icdm50108.2020.00031 · Cited by 230
“Therefore, we use the margin triplet loss [28] for model optimization so that positive and negative samples can be well-discriminated to some extent”
MethodBlind Backdoors in Deep Learning Models
Bagdasaryan, E., & Shmatikov, V. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2005.03823 · Cited by 94
“62% is high given the complexity of the face identification task, the fact that the model architecture and sampling [79] are not designed for identification, and the extreme imbalance of the training data.”
BackgroundRethinking Feature Distribution for Loss Functions in Image Classification
Wan, W., Zhong, Y., Li, T., & Chen, J. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00950 · Cited by 166
Dual Shared-Specific Multiview Subspace Clustering
Zhou, T., Zhang, C., Peng, X., Bhaskar, H., & Yang, J. (2020) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2019.2918495 · Cited by 171
Pose-Robust Face Recognition via Deep Residual Equivariant Mapping
Cao, K., Rong, Y., Li, C., Tang, X., & Loy, C. C. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00544 · Cited by 169
A review of building occupancy measurement systems
Sun, K., Zhao, Q., & Zou, J. (2020) · Energy and Buildings · doi:10.1016/j.enbuild.2020.109965 · Cited by 172
Sajjad, M., Ullah, F. U. M., Ullah, M., Christodoulou, G., Alaya Cheikh, F., Hijji, M., Muhammad, K., & Rodrigues, J. J. P. C. (2023) · Alexandria Engineering Journal · doi:10.1016/j.aej.2023.01.017 · Cited by 170
Few-Shot Learning Through an Information Retrieval Lens
Triantafillou, E., Zemel, R. S., & Urtasun, R. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1707.02610 · Cited by 62
“Example applications include signature verification [8] and face verification [9, 10].”
BackgroundMethodA Novel Image Classification Method with CNN-XGBoost Model
Ren, X., Guo, H., Li, S., Wang, S., & Li, J. (2017) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-64185-0_28 · Cited by 165
“XGBoost is widely applied in many domains [7, 8, 10] because of its high efficiency and accuracy.”
BackgroundCross-Modal Retrieval in the Cooking Context: Learning Semantic Text-Image Embeddings
Carvalho, M., Cadène, R., Picard, D., Soulier, L., Thome, N., & Cord, M. (2018) · The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval · doi:10.1145/3209978.3210036 · Cited by 171
RGB-IR Person Re-identification by Cross-Modality Similarity Preservation
Wu, A., Zheng, W.-S., Gong, S., & Lai, J. (2020) · International Journal of Computer Vision · doi:10.1007/s11263-019-01290-1 · Cited by 172
Face Image Quality Assessment: A Literature Survey
Schlett, T., Rathgeb, C., Henniger, O., Galbally, J., Fierrez, J., & Busch, C. (2022) · ACM Computing Surveys · doi:10.1145/3507901 · Cited by 170
“Evaluations used FaceNet [182] and ArcFace [127] for FR, and selected images using QSs from both SER-FIQ variants, FaceQnet v0 [53], an approach proposed by Best-Rowden in [172], three general IQAAs (BRISQUE [79], NIQE [80], PIQE [81]), as well as a COTS system (Neurotec Biometric SDK 11.1 [167]).”
MethodInfluentialOn Low-Resolution Face Recognition in the Wild: Comparisons and New Techniques
Li, P., Prieto, L., Mery, D., & Flynn, P. J. (2019) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2018.2890812 · Cited by 171
Kumar, R., Wang, W., Kumar, J., Yang, T., Khan, A., Ali, W., & Ali, I. (2021) · Computerized Medical Imaging and Graphics · doi:10.1016/j.compmedimag.2020.101812 · Cited by 172
Channel-level acceleration of deep face representations
Polyak, A., & Wolf, L. (2015) · IEEE Access · doi:10.1109/access.2015.2494536 · Cited by 171
“A recent work [10] further improves the training criterion by using triplet cost to increase the discriminability between identities.”
BackgroundMethodInfluentialRegion-Based Quality Estimation Network for Large-Scale Person Re-Identification
Song, G., Leng, B., Liu, Y., Hetang, C., & Cai, S. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.12305 · Cited by 166
KVQA: Knowledge-Aware Visual Question Answering
Shah, S., Mishra, A., Yadati, N., & Talukdar, P. P. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33018876 · Cited by 160
“Facenet is a deep architecture designed to obtain robust face representation.”
BackgroundMethodInfluentialG, V. K. B., Carneiro, G., & Reid, I. D. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1512.09272 · Cited by 171
“However, sampling all possible pairs or triplets from the training dataset can quickly become intractable, where the majority of those samples may produce small costs in (3)-(5), resulting in slow convergence [26].”
BackgroundMethodInfluentialMS-ASL: A Large-Scale Data Set and Benchmark for Understanding American Sign Language
Joze, H. R. V., & Koller, O. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1812.01053 · Cited by 142
Conditional Similarity Networks
Veit, A., Belongie, S., & Karaletsos, T. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.193 · Cited by 169
Deep Unified Model For Face Recognition Based on Convolution Neural Network and Edge Computing
Khan, M. Z., Harous, S., Hassan, S. U., Ghani Khan, M. U., Iqbal, R., & Mumtaz, S. (2019) · IEEE Access · doi:10.1109/access.2019.2918275 · Cited by 170
Dense Object Nets: Learning Dense Visual Object Descriptors By and For Robotic Manipulation
Florence, P., Manuelli, L., & Tedrake, R. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1806.08756 · Cited by 142
ShapeNet: A Shapelet-Neural Network Approach for Multivariate Time Series Classification
Li, G., Choi, B., Xu, J., S Bhowmick, S., Chun, K.-P., & Wong, G. L.-H. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i9.17018 · Cited by 169
“There have been several loss functions for unsupervised learning, such as word2vec (Mikolov et al. 2013), image similarity (Chechik et al. 2010), and face recognition (Schroff, Kalenichenko, and Philbin 2015).”
BackgroundRecent advances on loss functions in deep learning for computer vision
Tian, Y., Su, D., Lauria, S., & Liu, X. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.04.127 · Cited by 168
M2FNet: Multi-modal Fusion Network for Emotion Recognition in Conversation
Chudasama, V., Kar, P., Gudmalwar, A., Shah, N., Wasnik, P., & Onoe, N. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw56347.2022.00511 · Cited by 168
TristouNet: Triplet loss for speaker turn embedding
Bredin, H. (2017) · 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2017.7953194 · Cited by 168
“Though we do plan to test other triplet sampling strategies in the future, we chose to go with the one called “hard negative” in [8].”
BackgroundMethodInfluentialUnravelling Robustness of Deep Learning Based Face Recognition Against Adversarial Attacks
Goswami, G., Ratha, N., Agarwal, A., Singh, R., & Vatsa, M. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.12341 · Cited by 159
Content and Style Disentanglement for Artistic Style Transfer
Kotovenko, D., Sanakoyeu, A., Lang, S., & Ommer, B. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00452 · Cited by 165
“This resembles a triplet loss widely used in metric learning [25, 9].”
Method
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.