{"url":"/sota/facial-expression-recognition-on-raf-db","task":{"name":"Facial Expression Recognition (FER)","url":"/task/facial-expression-recognition","note":null},"dataset":{"name":"RAF-DB","url":"/dataset/raf-db"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"**Facial Expression Recognition (FER)** is a computer vision task aimed at identifying and categorizing emotional expressions depicted on a human face. The goal is to automate the process of determining emotions in real-time, by analyzing the various features of a face such as eyebrows, eyes, mouth, and other features, and mapping them to a set of emotions such as anger, fear, surprise, sadness and happiness.\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [DeXpression](https://arxiv.org/pdf/1509.05371v2.pdf) )</span>","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["Overall Accuracy","Avg. Accuracy"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Overall Accuracy":"higher","Avg. Accuracy":"higher"}},"counts":{"rows":35,"rows_with_code":23,"rows_with_paper_page":35,"rows_dated":35,"rows_using_additional_data":11},"rows":[{"rank_in_archive_order":1,"model":"ResEmoteNet","metrics":{"Overall Accuracy":"94.76"},"uses_additional_data":true,"paper_date":"2024-09-01","paper":"/paper/resemotenet-bridging-accuracy-and-loss","paper_url":"https://arxiv.org/abs/2409.10545v2","paper_title":"ResEmoteNet: Bridging Accuracy and Loss Reduction in Facial Emotion Recognition","code":"https://github.com/ArnabKumarRoy02/ResEmoteNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"FMAE","metrics":{"Overall Accuracy":"93.45"},"uses_additional_data":true,"paper_date":"2024-07-15","paper":"/paper/representation-learning-and-identity","paper_url":"https://arxiv.org/abs/2407.11243v2","paper_title":"Representation Learning and Identity Adversarial Training for Facial Behavior Understanding","code":"https://github.com/forever208/fmae-iat","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"QCS","metrics":{"Overall Accuracy":"93.02"},"uses_additional_data":false,"paper_date":"2024-11-04","paper":"/paper/qcs-feature-refining-from-quadruplet-cross","paper_url":"https://arxiv.org/abs/2411.01988v5","paper_title":"QCS: Feature Refining from Quadruplet Cross Similarity for Facial Expression Recognition","code":"https://github.com/birdwcp/qcs","n_code_links":1,"syntology":null},{"rank_in_archive_order":4,"model":"Norface","metrics":{"Overall Accuracy":"92.97"},"uses_additional_data":false,"paper_date":"2024-07-22","paper":"/paper/norface-improving-facial-expression-analysis","paper_url":"https://arxiv.org/abs/2407.15617v1","paper_title":"Norface: Improving Facial Expression Analysis by Identity Normalization","code":"https://github.com/liuhw01/Norface","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"S2D","metrics":{"Overall Accuracy":"92.57"},"uses_additional_data":false,"paper_date":"2023-12-09","paper":"/paper/from-static-to-dynamic-adapting-landmark-1","paper_url":"https://arxiv.org/abs/2312.05447v2","paper_title":"From Static to Dynamic: Adapting Landmark-Aware Image Models for Facial Expression Recognition in Videos","code":"https://github.com/msa-lmc/s2d","n_code_links":2,"syntology":{"n_ran":7,"n_unverified":7,"n_samples":14,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"BTN","metrics":{"Avg. Accuracy":"87.3","Overall Accuracy":"92.54"},"uses_additional_data":false,"paper_date":"2024-07-05","paper":"/paper/batch-transformer-look-for-attention-in-batch","paper_url":"https://arxiv.org/abs/2407.04218v1","paper_title":"Batch Transformer: Look for Attention in Batch","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":7,"model":"GReFEL","metrics":{"Overall Accuracy":"92.47"},"uses_additional_data":false,"paper_date":"2024-10-21","paper":"/paper/grefel-geometry-aware-reliable-facial","paper_url":"https://arxiv.org/abs/2410.15927v1","paper_title":"GReFEL: Geometry-Aware Reliable Facial Expression Learning under Bias and Imbalanced Data Distribution","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":8,"model":"DDAMFN++","metrics":{"Overall Accuracy":"92.34"},"uses_additional_data":true,"paper_date":"2023-08-25","paper":"/paper/a-dual-direction-attention-mixed-feature","paper_url":"https://scholar.google.com/citations?view_op=view_citation&hl=zh-CN&user=P4efBMcAAAAJ&citation_for_view=P4efBMcAAAAJ:d1gkVwhDpl0C","paper_title":"A Dual-Direction Attention Mixed Feature Network for Facial Expression Recognition","code":"https://github.com/simon20010923/DDAMFN","n_code_links":1,"syntology":null},{"rank_in_archive_order":9,"model":"DCJT","metrics":{"Overall Accuracy":"92.24"},"uses_additional_data":true,"paper_date":"2024-04-02","paper":"/paper/joint-training-on-multiple-datasets-with","paper_url":"https://ieeexplore.ieee.org/document/10483295","paper_title":"Joint Training on Multiple Datasets With Inconsistent Labeling Criteria for Facial Expression Recognition","code":"https://github.com/iiTzFrankie/DCJT","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"POSTER++","metrics":{"Overall Accuracy":"92.21"},"uses_additional_data":false,"paper_date":"2023-01-28","paper":"/paper/poster-v2-a-simpler-and-stronger-facial","paper_url":"https://arxiv.org/abs/2301.12149v2","paper_title":"POSTER++: A simpler and stronger facial expression recognition network","code":"https://github.com/talented-q/poster_v2","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"APViT","metrics":{"Overall Accuracy":"91.98"},"uses_additional_data":false,"paper_date":"2022-12-11","paper":"/paper/vision-transformer-with-attentive-pooling-for","paper_url":"https://arxiv.org/abs/2212.05463v1","paper_title":"Vision Transformer with Attentive Pooling for Robust Facial Expression Recognition","code":"https://github.com/youqingxiaozhua/apvit","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"DDAMFN","metrics":{"Overall Accuracy":"91.35"},"uses_additional_data":false,"paper_date":"2023-08-25","paper":"/paper/a-dual-direction-attention-mixed-feature","paper_url":"https://scholar.google.com/citations?view_op=view_citation&hl=zh-CN&user=P4efBMcAAAAJ&citation_for_view=P4efBMcAAAAJ:d1gkVwhDpl0C","paper_title":"A Dual-Direction Attention Mixed Feature Network for Facial Expression Recognition","code":"https://github.com/simon20010923/DDAMFN","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"ViT-base + MAE","metrics":{"Overall Accuracy":"91.07"},"uses_additional_data":false,"paper_date":"2022-07-22","paper":"/paper/facial-expression-recognition-using-vanilla","paper_url":"https://arxiv.org/abs/2207.11081v4","paper_title":"Emotion Separation and Recognition from a Facial Expression by Generating the Poker Face with Vision Transformers","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":14,"model":"LFNSB","metrics":{"Overall Accuracy":"91.07"},"uses_additional_data":false,"paper_date":"2024-08-01","paper":"/paper/a-lightweight-model-enhancing-facial","paper_url":"https://www.preprints.org/manuscript/202408.1304/v1","paper_title":"A Lightweight Model Enhancing Facial Expression Recognition with Spatial Bias and Cosine-Harmony Loss","code":"https://github.com/1chenchen22/LFNSB","n_code_links":1,"syntology":null},{"rank_in_archive_order":15,"model":"ExpLLM","metrics":{"Overall Accuracy":"91.03"},"uses_additional_data":false,"paper_date":"2024-09-04","paper":"/paper/expllm-towards-chain-of-thought-for-facial","paper_url":"https://arxiv.org/abs/2409.02828v1","paper_title":"ExpLLM: Towards Chain of Thought for Facial Expression Recognition","code":"https://github.com/starhiking/ExpLLM-TMM","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"EAC(ResNet-50)","metrics":{"Overall Accuracy":"90.35"},"uses_additional_data":false,"paper_date":"2022-07-21","paper":"/paper/learn-from-all-erasing-attention-consistency","paper_url":"https://arxiv.org/abs/2207.10299v2","paper_title":"Learn From All: Erasing Attention Consistency for Noisy Label Facial Expression Recognition","code":"https://github.com/zyh-uaiaaaa/erasing-attention-consistency","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":1}},{"rank_in_archive_order":17,"model":"Ada-DF","metrics":{"Overall Accuracy":"90.04"},"uses_additional_data":false,"paper_date":"2023-05-05","paper":"/paper/a-dual-branch-adaptive-distribution-fusion","paper_url":"https://ieeexplore.ieee.org/document/10097033","paper_title":"A Dual-Branch Adaptive Distribution Fusion Framework for Real-World Facial Expression Recognition","code":"https://github.com/taylor-xy0827/Ada-DF","n_code_links":1,"syntology":null},{"rank_in_archive_order":18,"model":"DAN","metrics":{"Overall Accuracy":"89.70"},"uses_additional_data":false,"paper_date":"2021-09-15","paper":"/paper/distract-your-attention-multi-head-cross","paper_url":"https://arxiv.org/abs/2109.07270v6","paper_title":"Distract Your Attention: Multi-head Cross Attention Network for Facial Expression Recognition","code":"https://github.com/yaoing/dan","n_code_links":2,"syntology":null},{"rank_in_archive_order":19,"model":"RUL (ResNet-18)","metrics":{"Overall Accuracy":"88.98"},"uses_additional_data":false,"paper_date":"2021-12-01","paper":"/paper/relative-uncertainty-learning-for-facial","paper_url":"http://proceedings.neurips.cc/paper/2021/hash/9332c513ef44b682e9347822c2e457ac-Abstract.html","paper_title":"Relative Uncertainty Learning for Facial Expression Recognition","code":"https://github.com/zyh-uaiaaaa/relative-uncertainty-learning","n_code_links":1,"syntology":null},{"rank_in_archive_order":20,"model":"PSR","metrics":{"Overall Accuracy":"88.98"},"uses_additional_data":true,"paper_date":"2020-07-17","paper":"/paper/pyramid-with-super-resolution-for-in-the-wild","paper_url":"https://doi.org/10.1109/ACCESS.2020.3010018","paper_title":"Pyramid With Super Resolution for In-the-Wild Facial Expression Recognition","code":"https://github.com/thanhhungqb/pyramid-super-resolution","n_code_links":1,"syntology":null},{"rank_in_archive_order":21,"model":"FerNeXt","metrics":{"Overall Accuracy":"88.56"},"uses_additional_data":false,"paper_date":"2023-10-20","paper":"/paper/fernext-facial-expression-recognition-using","paper_url":"https://ieeexplore.ieee.org/document/10278345","paper_title":"FerNeXt: Facial Expression Recognition Using ConvNeXt with Channel Attention","code":"https://github.com/OmarEl-Khashab/FerNeXt-Facial-Expression-Recognition-Using-ConvNeXt-with-Channel-Attention","n_code_links":1,"syntology":null},{"rank_in_archive_order":22,"model":"EfficientFace","metrics":{"Overall Accuracy":"88.36"},"uses_additional_data":false,"paper_date":"2021-05-18","paper":"/paper/robust-lightweight-facial-expression","paper_url":"https://ojs.aaai.org/index.php/AAAI/article/view/16465","paper_title":"Robust Lightweight Facial Expression Recognition Network with Label Distribution Training","code":"https://github.com/zengqunzhao/efficientface","n_code_links":1,"syntology":null},{"rank_in_archive_order":23,"model":"MA-Net","metrics":{"Overall Accuracy":"88.36"},"uses_additional_data":false,"paper_date":"2021-07-05","paper":"/paper/learning-deep-global-multi-scale-and-local","paper_url":"https://ieeexplore.ieee.org/document/9474949","paper_title":"Learning Deep Global Multi-scale and Local Attention Features for Facial Expression Recognition in the Wild","code":"https://github.com/zengqunzhao/ma-net","n_code_links":1,"syntology":null},{"rank_in_archive_order":24,"model":"DACL (ResNet-18)","metrics":{"Avg. Accuracy":"80.44","Overall Accuracy":"87.78"},"uses_additional_data":true,"paper_date":"2021-01-07","paper":"/paper/facial-expression-recognition-in-the-wild-via","paper_url":"https://openaccess.thecvf.com/content/WACV2021/html/Farzaneh_Facial_Expression_Recognition_in_the_Wild_via_Deep_Attentive_Center_WACV_2021_paper.html","paper_title":"Facial Expression Recognition in the Wild via Deep Attentive Center Loss","code":"https://github.com/amirhfarzaneh/dacl","n_code_links":1,"syntology":null},{"rank_in_archive_order":25,"model":"MixAugment","metrics":{"Avg. Accuracy":"77.30","Overall Accuracy":"87.54"},"uses_additional_data":false,"paper_date":"2022-05-09","paper":"/paper/mixaugment-mixup-augmentation-methods-for","paper_url":"https://arxiv.org/abs/2205.04442v1","paper_title":"MixAugment & Mixup: Augmentation Methods for Facial Expression Recognition","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":26,"model":"ViT-base","metrics":{"Overall Accuracy":"87.22"},"uses_additional_data":false,"paper_date":"2022-07-22","paper":"/paper/facial-expression-recognition-using-vanilla","paper_url":"https://arxiv.org/abs/2207.11081v4","paper_title":"Emotion Separation and Recognition from a Facial Expression by Generating the Poker Face with Vision Transformers","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":27,"model":"ViT-tiny","metrics":{"Overall Accuracy":"87.03"},"uses_additional_data":false,"paper_date":"2022-07-22","paper":"/paper/facial-expression-recognition-using-vanilla","paper_url":"https://arxiv.org/abs/2207.11081v4","paper_title":"Emotion Separation and Recognition from a Facial Expression by Generating the Poker Face with Vision Transformers","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":28,"model":"Ad-Corre","metrics":{"Overall Accuracy":"86.96"},"uses_additional_data":false,"paper_date":"2022-03-03","paper":"/paper/ad-corre-adaptive-correlation-based-loss-for","paper_url":"https://ieeexplore.ieee.org/document/9727163","paper_title":"Ad-Corre: Adaptive Correlation-Based Loss for Facial Expression Recognition in the Wild","code":"https://github.com/aliprf/Ad-Corre","n_code_links":1,"syntology":null},{"rank_in_archive_order":29,"model":"RAN (ResNet-18)","metrics":{"Overall Accuracy":"86.9"},"uses_additional_data":true,"paper_date":"2019-05-10","paper":"/paper/region-attention-networks-for-pose-and","paper_url":"https://arxiv.org/abs/1905.04075v2","paper_title":"Region Attention Networks for Pose and Occlusion Robust Facial Expression Recognition","code":"https://github.com/kaiwang960112/Challenge-condition-FER-dataset","n_code_links":1,"syntology":null},{"rank_in_archive_order":30,"model":"C-EXPR-NET","metrics":{"Avg. Accuracy":"87.5"},"uses_additional_data":false,"paper_date":"2023-01-01","paper":"/paper/multi-label-compound-expression-recognition-c","paper_url":"http://openaccess.thecvf.com//content/CVPR2023/html/Kollias_Multi-Label_Compound_Expression_Recognition_C-EXPR_Database__Network_CVPR_2023_paper.html","paper_title":"Multi-Label Compound Expression Recognition: C-EXPR Database & Network","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":31,"model":"C MT PSR","metrics":{"Avg. Accuracy":"84.8"},"uses_additional_data":true,"paper_date":"2024-01-02","paper":"/paper/distribution-matching-for-multi-task-learning","paper_url":"https://arxiv.org/abs/2401.01219v2","paper_title":"Distribution Matching for Multi-Task Learning of Classification Tasks: a Large-Scale Study on Faces & Beyond","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":32,"model":"C MT VGGFACE","metrics":{"Avg. Accuracy":"81.4"},"uses_additional_data":true,"paper_date":"2024-01-02","paper":"/paper/distribution-matching-for-multi-task-learning","paper_url":"https://arxiv.org/abs/2401.01219v2","paper_title":"Distribution Matching for Multi-Task Learning of Classification Tasks: a Large-Scale Study on Faces & Beyond","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":33,"model":"FaceBehaviorNet","metrics":{"Avg. Accuracy":"78"},"uses_additional_data":true,"paper_date":"2021-05-08","paper":"/paper/distribution-matching-for-heterogeneous-multi","paper_url":"https://arxiv.org/abs/2105.03790v1","paper_title":"Distribution Matching for Heterogeneous Multi-Task Learning: a Large-scale Face Study","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":34,"model":"VGG-FACE","metrics":{"Avg. Accuracy":"77.5"},"uses_additional_data":true,"paper_date":"2018-11-12","paper":"/paper/generating-faces-for-affect-analysis","paper_url":"https://arxiv.org/abs/1811.05027v2","paper_title":"Deep Neural Network Augmentation: Generating Faces for Affect Analysis","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":35,"model":"MT-ArcVGG","metrics":{"Avg. Accuracy":"76"},"uses_additional_data":false,"paper_date":"2019-09-25","paper":"/paper/expression-affect-action-unit-recognition-aff","paper_url":"https://arxiv.org/abs/1910.04855v1","paper_title":"Expression, Affect, Action Unit Recognition: Aff-Wild2, Multi-Task Learning and ArcFace","code":null,"n_code_links":0,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 7,081 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":7081,"papers_extracted_not_yet_verified":217,"boards_without_verdict":27,"papers_not_yet_extracted":2325},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":2,"rows_with_any_sample_ran":1,"distinct_papers_with_graph_line":2,"distinct_papers_with_any_sample_ran":1,"samples_over_distinct_papers":{"n_ran":7,"n_unverified":8,"n_samples":15,"n_pointer_only_licence":1,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":7,"n_unverified":8,"n_samples":15,"n_pointer_only_licence":1,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}