{"url":"/sota/lesion-segmentation-on-isic-2018","task":{"name":"Lesion Segmentation","url":"/task/lesion-segmentation","note":null},"dataset":{"name":"ISIC 2018","url":null},"category":"Computer Vision","categories":["Computer Vision","Medical"],"category_note":null,"description":"Lesion segmentation is the task of segmenting out lesions from other objects in medical based images.\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [D-UNet](https://arxiv.org/pdf/1908.05104v1.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":["mean Dice","F1-Score","Mean IoU"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"mean Dice":"higher","F1-Score":"higher","Mean IoU":"higher"}},"counts":{"rows":17,"rows_with_code":15,"rows_with_paper_page":17,"rows_dated":17,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"Polar Res-U-Net++","metrics":{"mean Dice":"0.9253"},"uses_additional_data":false,"paper_date":"2021-09-29","paper":"/paper/training-on-polar-image-transformations","paper_url":"https://ieeexplore.ieee.org/document/9551998","paper_title":"Training on Polar Image Transformations Improves Biomedical Image Segmentation","code":"https://github.com/marinbenc/medical-polar-training","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"DuAT","metrics":{"Mean IoU":"0.867","mean Dice":"0.923"},"uses_additional_data":false,"paper_date":"2022-12-21","paper":"/paper/duat-dual-aggregation-transformer-network-for","paper_url":"https://arxiv.org/abs/2212.11677v1","paper_title":"DuAT: Dual-Aggregation Transformer Network for Medical Image Segmentation","code":"https://github.com/Barrett-python/DuAT","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"ProMISe","metrics":{"Mean IoU":"0.850","mean Dice":"0.921"},"uses_additional_data":false,"paper_date":"2024-03-07","paper":"/paper/promise-promptable-medical-image-segmentation","paper_url":"https://arxiv.org/abs/2403.04164v3","paper_title":"ProMISe: Promptable Medical Image Segmentation using SAM","code":"https://github.com/xinkunwang111/promise","n_code_links":1,"syntology":null},{"rank_in_archive_order":4,"model":"RMSM UNet + DF-RAM +EF-RAM","metrics":{"mean Dice":"0.9152"},"uses_additional_data":false,"paper_date":"2021-11-16","paper":"/paper/exploring-dual-attention-mechanism-with-multi","paper_url":"https://arxiv.org/abs/2111.08708v3","paper_title":"Automated skin lesion segmentation using multi-scale feature extraction scheme and dual-attention mechanism","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":5,"model":"BAT","metrics":{"Mean IoU":"0.843","mean Dice":"0.912"},"uses_additional_data":false,"paper_date":"2021-10-08","paper":"/paper/boundary-aware-transformers-for-skin-lesion","paper_url":"https://arxiv.org/abs/2110.03864v1","paper_title":"Boundary-aware Transformers for Skin Lesion Segmentation","code":"https://github.com/jcwang123/BA-Transformer","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":1}},{"rank_in_archive_order":6,"model":"SegMed","metrics":{"Mean IoU":"0.841","mean Dice":"0.911"},"uses_additional_data":false,"paper_date":"2025-01-31","paper":"/paper/from-semantic-segmentation-of-natural-images","paper_url":"https://link.springer.com/chapter/10.1007/978-3-031-80507-3_12","paper_title":"From Semantic Segmentation of Natural Images to Medical Image Segmentation Using ViT-Based Architectures","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":7,"model":"MobileUNETR","metrics":{"Mean IoU":"0.8456","mean Dice":"0.9074"},"uses_additional_data":false,"paper_date":"2024-09-04","paper":"/paper/mobileunetr-a-lightweight-end-to-end-hybrid","paper_url":"https://arxiv.org/abs/2409.03062v1","paper_title":"MobileUNETR: A Lightweight End-To-End Hybrid Vision Transformer For Efficient Medical Image Segmentation","code":"https://github.com/osupcvlab/mobileunetr","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"DermoSegDiff-A","metrics":{"mean Dice":"0.9005"},"uses_additional_data":false,"paper_date":"2023-08-05","paper":"/paper/dermosegdiff-a-boundary-aware-segmentation","paper_url":"https://arxiv.org/abs/2308.02959v1","paper_title":"DermoSegDiff: A Boundary-aware Segmentation Diffusion Model for Skin Lesion Delineation","code":"https://github.com/mindflow-institue/dermosegdiff","n_code_links":1,"syntology":null},{"rank_in_archive_order":9,"model":"DoubleU-Net","metrics":{"mean Dice":"0.8962"},"uses_additional_data":false,"paper_date":"2020-06-08","paper":"/paper/doubleu-net-a-deep-convolutional-neural","paper_url":"https://arxiv.org/abs/2006.04868v2","paper_title":"DoubleU-Net: A Deep Convolutional Neural Network for Medical Image Segmentation","code":"https://github.com/DebeshJha/2020-CBMS-DoubleU-Net","n_code_links":4,"syntology":{"n_ran":2,"n_unverified":3,"n_samples":5,"n_pointer_only_licence":5}},{"rank_in_archive_order":10,"model":"MCGU-Net","metrics":{"mean Dice":"0.895"},"uses_additional_data":false,"paper_date":"2020-03-10","paper":"/paper/multi-level-context-gating-of-embedded","paper_url":"https://arxiv.org/abs/2003.05056v1","paper_title":"Multi-level Context Gating of Embedded Collective Knowledge for Medical Image Segmentation","code":"https://github.com/rezazad68/BCDU-Net","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"MSRF-Net","metrics":{"mean Dice":"0.8813"},"uses_additional_data":false,"paper_date":"2021-05-16","paper":"/paper/msrf-net-a-multi-scale-residual-fusion","paper_url":"https://arxiv.org/abs/2105.07451v2","paper_title":"MSRF-Net: A Multi-Scale Residual Fusion Network for Biomedical Image Segmentation","code":"https://github.com/NoviceMAn-prog/MSRF-Net","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"Attn U-Net + Multi-Input + FTL","metrics":{"mean Dice":"0.856"},"uses_additional_data":false,"paper_date":"2018-10-18","paper":"/paper/a-novel-focal-tversky-loss-function-with","paper_url":"http://arxiv.org/abs/1810.07842v1","paper_title":"A Novel Focal Tversky loss function with improved Attention U-Net for lesion segmentation","code":"https://github.com/nabsabraham/focal-tversky-unet","n_code_links":6,"syntology":null},{"rank_in_archive_order":13,"model":"AIM++ (256x256, 1.5m parameters, 10% labeled data, no pretraining)","metrics":{"mean Dice":"0.85"},"uses_additional_data":false,"paper_date":"2024-01-25","paper":"/paper/inconsistency-masks-removing-the-uncertainty","paper_url":"https://arxiv.org/abs/2401.14387v2","paper_title":"Inconsistency Masks: Removing the Uncertainty from Input-Pseudo-Label Pairs","code":"https://github.com/michaelvorndran/inconsistencymasks","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"BCDU-net","metrics":{"mean Dice":"0.847"},"uses_additional_data":false,"paper_date":"2019-08-31","paper":"/paper/bi-directional-convlstm-u-net-with-densley","paper_url":"https://arxiv.org/abs/1909.00166v1","paper_title":"Bi-Directional ConvLSTM U-Net with Densley Connected Convolutions","code":"https://github.com/rezazad68/BCDU-Net","n_code_links":4,"syntology":null},{"rank_in_archive_order":15,"model":"U-Net + FTL","metrics":{"mean Dice":"0.829"},"uses_additional_data":false,"paper_date":"2018-10-18","paper":"/paper/a-novel-focal-tversky-loss-function-with","paper_url":"http://arxiv.org/abs/1810.07842v1","paper_title":"A Novel Focal Tversky loss function with improved Attention U-Net for lesion segmentation","code":"https://github.com/nabsabraham/focal-tversky-unet","n_code_links":6,"syntology":null},{"rank_in_archive_order":16,"model":"Attn U-Net + DL","metrics":{"mean Dice":"0.806"},"uses_additional_data":false,"paper_date":"2018-10-18","paper":"/paper/a-novel-focal-tversky-loss-function-with","paper_url":"http://arxiv.org/abs/1810.07842v1","paper_title":"A Novel Focal Tversky loss function with improved Attention U-Net for lesion segmentation","code":"https://github.com/nabsabraham/focal-tversky-unet","n_code_links":6,"syntology":null},{"rank_in_archive_order":17,"model":"BCDU-Net (d=3)","metrics":{"F1-Score":"0.851"},"uses_additional_data":false,"paper_date":"2019-08-31","paper":"/paper/bi-directional-convlstm-u-net-with-densley","paper_url":"https://arxiv.org/abs/1909.00166v1","paper_title":"Bi-Directional ConvLSTM U-Net with Densley Connected Convolutions","code":"https://github.com/rezazad68/BCDU-Net","n_code_links":4,"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 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":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":2,"n_unverified":4,"n_samples":6,"n_pointer_only_licence":6,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":2,"n_unverified":4,"n_samples":6,"n_pointer_only_licence":6,"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"}}}