{"url":"/sota/object-counting-on-fsc147","task":{"name":"Object Counting","url":"/task/object-counting","note":null},"dataset":{"name":"FSC147","url":"/dataset/fsc147"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"The goal of **Object Counting** task is to count the number of object instances in a single image or video sequence. It has many real-world applications such as traffic flow monitoring, crowdedness estimation, and product counting.\n\n\n<span class=\"description-source\">Source: [Learning to Count Objects with Few Exemplar Annotations ](https://arxiv.org/abs/1905.07898)</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":["MAE(test)","MAE(val)","RMSE(test)","RMSE(val)"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"MAE(test)":"lower","MAE(val)":"lower","RMSE(test)":"lower","RMSE(val)":"lower"}},"counts":{"rows":19,"rows_with_code":16,"rows_with_paper_page":19,"rows_dated":19,"rows_using_additional_data":2},"rows":[{"rank_in_archive_order":1,"model":"CountGD","metrics":{"MAE(test)":"5.74","MAE(val)":"7.1","RMSE(test)":"24.09","RMSE(val)":"26.08"},"uses_additional_data":false,"paper_date":"2024-07-05","paper":"/paper/countgd-multi-modal-open-world-counting","paper_url":"https://arxiv.org/abs/2407.04619v1","paper_title":"CountGD: Multi-Modal Open-World Counting","code":"https://github.com/niki-amini-naieni/CountGD","n_code_links":2,"syntology":{"n_ran":4,"n_unverified":5,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"GeCo","metrics":{"MAE(test)":"7.91","MAE(val)":"9.52","RMSE(test)":"54.28","RMSE(val)":"43.00"},"uses_additional_data":false,"paper_date":"2024-09-27","paper":"/paper/a-novel-unified-architecture-for-low-shot","paper_url":"https://arxiv.org/abs/2409.18686v2","paper_title":"A Novel Unified Architecture for Low-Shot Counting by Detection and Segmentation","code":"https://github.com/jerpelhan/GeCo","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"DAVE","metrics":{"MAE(test)":"8.66","MAE(val)":"8.91","RMSE(test)":"32.36","RMSE(val)":"28.08"},"uses_additional_data":false,"paper_date":"2024-04-25","paper":"/paper/dave-a-detect-and-verify-paradigm-for-low","paper_url":"https://arxiv.org/abs/2404.16622v1","paper_title":"DAVE -- A Detect-and-Verify Paradigm for Low-Shot Counting","code":"https://github.com/jerpelhan/dave","n_code_links":1,"syntology":{"n_ran":16,"n_unverified":2,"n_samples":18,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"CACViT","metrics":{"MAE(test)":"9.13","MAE(val)":"10.63","RMSE(test)":"48.96","RMSE(val)":"37.95"},"uses_additional_data":false,"paper_date":"2023-05-08","paper":"/paper/vision-transformer-off-the-shelf-a-surprising","paper_url":"https://arxiv.org/abs/2305.04440v2","paper_title":"Vision Transformer Off-the-Shelf: A Surprising Baseline for Few-Shot Class-Agnostic Counting","code":"https://github.com/Xu3XiWang/CACViT","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"SSD","metrics":{"MAE(test)":"9.58","MAE(val)":"9.73","RMSE(test)":"64.13","RMSE(val)":"29.72"},"uses_additional_data":false,"paper_date":"2024-05-20","paper":"/paper/learning-spatial-similarity-distribution-for","paper_url":"https://arxiv.org/abs/2405.11770v1","paper_title":"Learning Spatial Similarity Distribution for Few-shot Object Counting","code":"https://github.com/CBalance/SSD","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":6,"n_samples":10,"n_pointer_only_licence":10}},{"rank_in_archive_order":6,"model":"LOCA","metrics":{"MAE(test)":"10.79","MAE(val)":"10.24","RMSE(test)":"56.97","RMSE(val)":"32.56"},"uses_additional_data":false,"paper_date":"2022-11-15","paper":"/paper/a-low-shot-object-counting-network-with","paper_url":"https://arxiv.org/abs/2211.08217v2","paper_title":"A Low-Shot Object Counting Network With Iterative Prototype Adaptation","code":"https://github.com/djukicn/loca","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":2,"n_samples":2,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"SemAug-CounTR","metrics":{"MAE(test)":"11.32","MAE(val)":"12.31","RMSE(test)":"77.50","RMSE(val)":"41.65"},"uses_additional_data":true,"paper_date":"2023-10-26","paper":"/paper/semantic-generative-augmentations-for-few","paper_url":"https://arxiv.org/abs/2311.16122v1","paper_title":"Semantic Generative Augmentations for Few-Shot Counting","code":"https://github.com/perladoubinsky/SemAug","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"CounTR","metrics":{"MAE(test)":"11.95","MAE(val)":"13.13","RMSE(test)":"91.23","RMSE(val)":"49.83"},"uses_additional_data":false,"paper_date":"2022-08-29","paper":"/paper/countr-transformer-based-generalised-visual","paper_url":"https://arxiv.org/abs/2208.13721v3","paper_title":"CounTR: Transformer-based Generalised Visual Counting","code":"https://github.com/Verg-Avesta/CounTR","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":3,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"SemAug-SAFECount","metrics":{"MAE(test)":"12.74","MAE(val)":"12.59","RMSE(test)":"89.90","RMSE(val)":"44.95"},"uses_additional_data":true,"paper_date":"2023-10-26","paper":"/paper/semantic-generative-augmentations-for-few","paper_url":"https://arxiv.org/abs/2311.16122v1","paper_title":"Semantic Generative Augmentations for Few-Shot Counting","code":"https://github.com/perladoubinsky/SemAug","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"SPDCN","metrics":{"MAE(test)":"13.51","MAE(val)":"14.59","RMSE(test)":"96.80","RMSE(val)":"49.97"},"uses_additional_data":false,"paper_date":"2022-12-21","paper":"/paper/scale-prior-deformable-convolution-for","paper_url":"https://bmvc2022.mpi-inf.mpg.de/313/","paper_title":"Scale-Prior Deformable Convolution for Exemplar-Guided Class-Agnostic Counting","code":"https://github.com/Elin24/SPDCN-CAC","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"GCA-SUN","metrics":{"MAE(test)":"14.00","MAE(val)":"16.06","RMSE(test)":"92.19","RMSE(val)":"53.04"},"uses_additional_data":false,"paper_date":"2024-09-18","paper":"/paper/gca-sun-a-gated-context-aware-swin-unet-for","paper_url":"https://arxiv.org/abs/2409.12249v2","paper_title":"GCA-SUNet: A Gated Context-Aware Swin-UNet for Exemplar-Free Counting","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":12,"model":"SAFECount","metrics":{"MAE(test)":"14.32","MAE(val)":"15.28","RMSE(test)":"85.54","RMSE(val)":"47.20"},"uses_additional_data":false,"paper_date":"2022-01-22","paper":"/paper/iterative-correlation-based-feature","paper_url":"https://arxiv.org/abs/2201.08959v5","paper_title":"Few-shot Object Counting with Similarity-Aware Feature Enhancement","code":"https://github.com/zhiyuanyou/SAFECount","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"BMNet+","metrics":{"MAE(test)":"14.62","MAE(val)":"15.74","RMSE(test)":"91.83","RMSE(val)":"58.53"},"uses_additional_data":false,"paper_date":"2022-03-16","paper":"/paper/represent-compare-and-learn-a-similarity","paper_url":"https://arxiv.org/abs/2203.08354v1","paper_title":"Represent, Compare, and Learn: A Similarity-Aware Framework for Class-Agnostic Counting","code":"https://github.com/flyinglynx/Bilinear-Matching-Network","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":14,"model":"LaoNet","metrics":{"MAE(test)":"15.78","MAE(val)":"17.11","RMSE(test)":"97.15","RMSE(val)":"56.81"},"uses_additional_data":false,"paper_date":"2021-12-11","paper":"/paper/object-counting-you-only-need-to-look-at-one","paper_url":"https://arxiv.org/abs/2112.05993v1","paper_title":"Object Counting: You Only Need to Look at One","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":15,"model":"CounTX (uses text descriptions instead of visual exemplars)","metrics":{"MAE(test)":"15.88","MAE(val)":"17.10","RMSE(test)":"106.29","RMSE(val)":"65.61"},"uses_additional_data":false,"paper_date":"2023-06-02","paper":"/paper/open-world-text-specified-object-counting","paper_url":"https://arxiv.org/abs/2306.01851v2","paper_title":"Open-world Text-specified Object Counting","code":"https://github.com/niki-amini-naieni/countx","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"Counting-DETR","metrics":{"MAE(test)":"16.79","RMSE(test)":"123.56"},"uses_additional_data":false,"paper_date":"2022-07-22","paper":"/paper/few-shot-object-counting-and-detection","paper_url":"https://arxiv.org/abs/2207.10988v2","paper_title":"Few-shot Object Counting and Detection","code":"https://github.com/vinairesearch/counting-detr","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"RCC","metrics":{"MAE(test)":"17.12","MAE(val)":"17.49","RMSE(test)":"104.53","RMSE(val)":"58.81"},"uses_additional_data":false,"paper_date":"2022-05-20","paper":"/paper/learning-to-count-anything-reference-less","paper_url":"https://arxiv.org/abs/2205.10203v2","paper_title":"Learning to Count Anything: Reference-less Class-agnostic Counting with Weak Supervision","code":"https://github.com/SinicaGroup/Class-agnostic-Few-shot-Object-Counting","n_code_links":2,"syntology":{"n_ran":1,"n_unverified":4,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":18,"model":"Omnicount (Open vocabulary, multi-label, without training)","metrics":{"MAE(test)":"18.63","RMSE(test)":"112"},"uses_additional_data":false,"paper_date":"2024-03-08","paper":"/paper/omnicount-multi-label-object-counting-with","paper_url":"https://arxiv.org/abs/2403.05435v8","paper_title":"OmniCount: Multi-label Object Counting with Semantic-Geometric Priors","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":19,"model":"FamNet","metrics":{"MAE(test)":"22.08","MAE(val)":"23.75","RMSE(test)":"99.54","RMSE(val)":"69.07"},"uses_additional_data":false,"paper_date":"2021-04-16","paper":"/paper/learning-to-count-everything","paper_url":"https://arxiv.org/abs/2104.08391v1","paper_title":"Learning To Count Everything","code":"https://github.com/cvlab-stonybrook/LearningToCountEverything","n_code_links":1,"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":5,"distinct_papers_with_graph_line":9,"distinct_papers_with_any_sample_ran":5,"samples_over_distinct_papers":{"n_ran":26,"n_unverified":28,"n_samples":54,"n_pointer_only_licence":10,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":26,"n_unverified":28,"n_samples":54,"n_pointer_only_licence":10,"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"}}}