{"url":"/sota/image-clustering-on-imagenet-dog-15","task":{"name":"Image Clustering","url":"/task/image-clustering","note":null},"dataset":{"name":"Imagenet-dog-15","url":null},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"Models that partition the dataset into semantically meaningful clusters without having access to the ground truth labels. \r\n\r\n<span style=\"color:grey; opacity: 0.6\"> Image credit: ImageNet clustering results of [SCAN: Learning to Classify Images without Labels (ECCV 2020)](https://arxiv.org/abs/2005.12320) </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":["Accuracy","NMI","ARI","Backbone","Image Size"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Accuracy":"higher","NMI":null,"ARI":null,"Backbone":null,"Image Size":null}},"counts":{"rows":20,"rows_with_code":20,"rows_with_paper_page":20,"rows_dated":20,"rows_using_additional_data":3},"rows":[{"rank_in_archive_order":1,"model":"MAE-CT (best)","metrics":{"ARI":"0.879","Accuracy":"0.943","Backbone":"ViT-H/16","Image Size":"224","NMI":"0.904"},"uses_additional_data":true,"paper_date":"2023-04-20","paper":"/paper/contrastive-tuning-a-little-help-to-make","paper_url":"https://arxiv.org/abs/2304.10520v2","paper_title":"Contrastive Tuning: A Little Help to Make Masked Autoencoders Forget","code":"https://github.com/ml-jku/mae-ct","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"MAE-CT (mean)","metrics":{"ARI":"0.821","Accuracy":"0.874","Backbone":"ViT-H/16","Image Size":"224","NMI":"0.882"},"uses_additional_data":true,"paper_date":"2023-04-20","paper":"/paper/contrastive-tuning-a-little-help-to-make","paper_url":"https://arxiv.org/abs/2304.10520v2","paper_title":"Contrastive Tuning: A Little Help to Make Masked Autoencoders Forget","code":"https://github.com/ml-jku/mae-ct","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"PRO-DSC","metrics":{"Accuracy":"0.840","NMI":"0.812"},"uses_additional_data":true,"paper_date":"2025-03-21","paper":"/paper/exploring-a-principled-framework-for-deep-1","paper_url":"https://arxiv.org/abs/2503.17288v1","paper_title":"Exploring a Principled Framework for Deep Subspace Clustering","code":"https://github.com/mengxianghan123/PRO-DSC","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":0,"n_samples":4,"n_pointer_only_licence":4}},{"rank_in_archive_order":4,"model":"ProPos*","metrics":{"ARI":"0.675","Accuracy":"0.775","Backbone":"ResNet-34","Image Size":"224","NMI":"0.737"},"uses_additional_data":false,"paper_date":"2021-11-23","paper":"/paper/exploring-non-contrastive-representation-1","paper_url":"https://arxiv.org/abs/2111.11821v2","paper_title":"Learning Representation for Clustering via Prototype Scattering and Positive Sampling","code":"https://github.com/hzzone/propos","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"ProPos","metrics":{"ARI":"0.627","Accuracy":"0.745","Backbone":"ResNet-34","Image Size":"96","NMI":"0.692"},"uses_additional_data":false,"paper_date":"2021-11-23","paper":"/paper/exploring-non-contrastive-representation-1","paper_url":"https://arxiv.org/abs/2111.11821v2","paper_title":"Learning Representation for Clustering via Prototype Scattering and Positive Sampling","code":"https://github.com/hzzone/propos","n_code_links":1,"syntology":null},{"rank_in_archive_order":6,"model":"DPAC","metrics":{"ARI":"0.598","Accuracy":"0.726","Backbone":"ResNet-34","NMI":"0.667"},"uses_additional_data":false,"paper_date":"2024-07-07","paper":"/paper/deep-probability-aggregation-clustering","paper_url":"https://arxiv.org/abs/2407.05246v2","paper_title":"Deep Online Probability Aggregation Clustering","code":"https://github.com/aomandechenai/deep-probability-aggregation-clustering","n_code_links":1,"syntology":{"n_ran":5,"n_unverified":5,"n_samples":10,"n_pointer_only_licence":10}},{"rank_in_archive_order":7,"model":"ConCURL","metrics":{"ARI":"0.531","Accuracy":"0.695","NMI":"0.63"},"uses_additional_data":false,"paper_date":"2021-05-04","paper":"/paper/representation-learning-for-clustering-via","paper_url":"https://arxiv.org/abs/2105.01289v2","paper_title":"Representation Learning for Clustering via Building Consensus","code":"https://github.com/JayanthRR/ConCURL_NCE","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":6,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":8,"model":"SPICE","metrics":{"ARI":"0.526","Accuracy":"0.675","Backbone":"ResNet-34","NMI":"0.627"},"uses_additional_data":false,"paper_date":"2021-03-17","paper":"/paper/spice-semantic-pseudo-labeling-for-image","paper_url":"https://arxiv.org/abs/2103.09382v3","paper_title":"SPICE: Semantic Pseudo-labeling for Image Clustering","code":"https://github.com/niuchuangnn/SPICE","n_code_links":1,"syntology":null},{"rank_in_archive_order":9,"model":"TCL","metrics":{"ARI":"0.516","Accuracy":"0.644","NMI":"0.623"},"uses_additional_data":false,"paper_date":"2022-10-21","paper":"/paper/twin-contrastive-learning-for-online","paper_url":"https://arxiv.org/abs/2210.11680v1","paper_title":"Twin Contrastive Learning for Online Clustering","code":"https://github.com/Yunfan-Li/Twin-Contrastive-Learning","n_code_links":2,"syntology":null},{"rank_in_archive_order":10,"model":"IDFD","metrics":{"ARI":"0.413","Accuracy":"0.591","Image Size":"96","NMI":"0.546"},"uses_additional_data":false,"paper_date":"2021-05-31","paper":"/paper/clustering-friendly-representation-learning-1","paper_url":"https://arxiv.org/abs/2106.00131v1","paper_title":"Clustering-friendly Representation Learning via Instance Discrimination and Feature Decorrelation","code":"https://github.com/TTN-YKK/Clustering_friendly_representation_learning","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"MiCE","metrics":{"ARI":"0.286","Accuracy":"0.439","Image Size":"96","NMI":"0.423"},"uses_additional_data":false,"paper_date":"2021-05-05","paper":"/paper/mice-mixture-of-contrastive-experts-for-1","paper_url":"https://arxiv.org/abs/2105.01899v1","paper_title":"MiCE: Mixture of Contrastive Experts for Unsupervised Image Clustering","code":"https://github.com/TsungWeiTsai/MiCE","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":2,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":12,"model":"C3","metrics":{"ARI":"0.28","Accuracy":"0.434","NMI":"0.448"},"uses_additional_data":false,"paper_date":"2022-11-14","paper":"/paper/c3-cross-instance-guided-contrastive","paper_url":"https://arxiv.org/abs/2211.07136v4","paper_title":"C3: Cross-instance guided Contrastive Clustering","code":"https://github.com/Armanfard-Lab/C3","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"CC","metrics":{"ARI":"0.274","Accuracy":"0.429","Image Size":"224","NMI":"0.445"},"uses_additional_data":false,"paper_date":"2020-09-21","paper":"/paper/contrastive-clustering","paper_url":"https://arxiv.org/abs/2009.09687v1","paper_title":"Contrastive Clustering","code":"https://github.com/Yunfan-Li/Contrastive-Clustering","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":5,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":14,"model":"DCCM","metrics":{"Accuracy":"0.383","NMI":"0.321"},"uses_additional_data":false,"paper_date":"2019-04-15","paper":"/paper/deep-comprehensive-correlation-mining-for","paper_url":"https://arxiv.org/abs/1904.06925v3","paper_title":"Deep Comprehensive Correlation Mining for Image Clustering","code":"https://github.com/Cory-M/DCCM","n_code_links":1,"syntology":null},{"rank_in_archive_order":15,"model":"CoHiClust","metrics":{"ARI":"0.232","Accuracy":"0.355","Backbone":"ResNet-50","NMI":"0.411"},"uses_additional_data":false,"paper_date":"2023-03-03","paper":"/paper/contrastive-hierarchical-clustering","paper_url":"https://arxiv.org/abs/2303.03389v2","paper_title":"Contrastive Hierarchical Clustering","code":"https://github.com/michalznalezniak/contrastive-hierarchical-clustering","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"DAC","metrics":{"Accuracy":"0.275","NMI":"0.219"},"uses_additional_data":false,"paper_date":"2017-10-01","paper":"/paper/deep-adaptive-image-clustering","paper_url":"http://openaccess.thecvf.com/content_iccv_2017/html/Chang_Deep_Adaptive_Image_ICCV_2017_paper.html","paper_title":"Deep Adaptive Image Clustering","code":"https://github.com/vector-1127/DAC","n_code_links":1,"syntology":null},{"rank_in_archive_order":17,"model":"DEC","metrics":{"Accuracy":"0.195","NMI":"0.122"},"uses_additional_data":false,"paper_date":"2015-11-19","paper":"/paper/unsupervised-deep-embedding-for-clustering","paper_url":"http://arxiv.org/abs/1511.06335v2","paper_title":"Unsupervised Deep Embedding for Clustering Analysis","code":"https://github.com/piiswrong/dec","n_code_links":23,"syntology":{"n_ran":2,"n_unverified":23,"n_samples":25,"n_pointer_only_licence":3}},{"rank_in_archive_order":18,"model":"VAE","metrics":{"Accuracy":"0.179","NMI":"0.107"},"uses_additional_data":false,"paper_date":"2013-12-20","paper":"/paper/auto-encoding-variational-bayes","paper_url":"http://arxiv.org/abs/1312.6114v10","paper_title":"Auto-Encoding Variational Bayes","code":"https://github.com/microsoft/recommenders","n_code_links":144,"syntology":{"n_ran":112,"n_unverified":87,"n_samples":199,"n_pointer_only_licence":103}},{"rank_in_archive_order":19,"model":"GAN","metrics":{"Accuracy":"0.174","NMI":"0.121"},"uses_additional_data":false,"paper_date":"2015-11-19","paper":"/paper/unsupervised-representation-learning-with-1","paper_url":"http://arxiv.org/abs/1511.06434v2","paper_title":"Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks","code":"https://github.com/tensorflow/models/blob/master/research/slim/nets/dcgan.py","n_code_links":258,"syntology":{"n_ran":113,"n_unverified":106,"n_samples":219,"n_pointer_only_licence":111}},{"rank_in_archive_order":20,"model":"JULE","metrics":{"Accuracy":"0.138","NMI":"0.054"},"uses_additional_data":false,"paper_date":"2016-04-13","paper":"/paper/joint-unsupervised-learning-of-deep","paper_url":"http://arxiv.org/abs/1604.03628v3","paper_title":"Joint Unsupervised Learning of Deep Representations and Image Clusters","code":"https://github.com/jwyang/jule.torch","n_code_links":3,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}}],"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 6,264 of the 9,581 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":9581,"papers_checked":6264,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":3316},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-24T18:15:14+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":9,"rows_with_any_sample_ran":8,"distinct_papers_with_graph_line":9,"distinct_papers_with_any_sample_ran":8,"samples_over_distinct_papers":{"n_ran":240,"n_unverified":235,"n_samples":475,"n_pointer_only_licence":234,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":240,"n_unverified":235,"n_samples":475,"n_pointer_only_licence":234,"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"}}}