{"url":"/sota/3d-object-detection-on-scannetv2","task":{"name":"3D Object Detection","url":"/task/3d-object-detection","note":null},"dataset":{"name":"ScanNetV2","url":"/dataset/scannet"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"**3D Object Detection** is a task in computer vision where the goal is to identify and locate objects in a 3D environment based on their shape, location, and orientation. It involves detecting the presence of objects and determining their location in the 3D space in real-time. This task is crucial for applications such as autonomous vehicles, robotics, and augmented reality.\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [AVOD](https://github.com/kujason/avod) )</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":["mAP@0.5","mAP@0.25"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"mAP@0.5":"higher","mAP@0.25":"higher"}},"counts":{"rows":33,"rows_with_code":28,"rows_with_paper_page":32,"rows_dated":32,"rows_using_additional_data":2},"rows":[{"rank_in_archive_order":1,"model":"DEST (based on V-DETR) (TTA)","metrics":{"mAP@0.25":"78.8","mAP@0.5":"67.9"},"uses_additional_data":false,"paper_date":"2025-03-18","paper":"/paper/state-space-model-meets-transformer-a-new-1","paper_url":"https://arxiv.org/abs/2503.14493v1","paper_title":"State Space Model Meets Transformer: A New Paradigm for 3D Object Detection","code":"https://github.com/OpenSpaceAI/DEST3D","n_code_links":1,"syntology":{"n_ran":6,"n_unverified":4,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"UniDet3D","metrics":{"mAP@0.25":"77.5","mAP@0.5":"66.1"},"uses_additional_data":false,"paper_date":"2024-09-06","paper":"/paper/unidet3d-multi-dataset-indoor-3d-object","paper_url":"https://arxiv.org/abs/2409.04234v1","paper_title":"UniDet3D: Multi-dataset Indoor 3D Object Detection","code":"https://github.com/filapro/unidet3d","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"V-DETR","metrics":{"mAP@0.25":"77.8","mAP@0.5":"65.9"},"uses_additional_data":false,"paper_date":"2023-08-08","paper":"/paper/v-detr-detr-with-vertex-relative-position","paper_url":"https://arxiv.org/abs/2308.04409v1","paper_title":"V-DETR: DETR with Vertex Relative Position Encoding for 3D Object Detection","code":"https://github.com/yichaoshen-ms/v-detr","n_code_links":1,"syntology":{"n_ran":10,"n_unverified":2,"n_samples":12,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"UDeerMvDETR","metrics":{"mAP@0.25":"78.0","mAP@0.5":"65.8"},"uses_additional_data":false,"paper_date":"2024-08-13","paper":"/paper/mv-detr-multi-modality-indoor-object","paper_url":"https://arxiv.org/abs/2408.06604v1","paper_title":"MV-DETR: Multi-modality indoor object detection by Multi-View DEtecton TRansformers","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":5,"model":"OneFormer3D","metrics":{"mAP@0.25":"76.9","mAP@0.5":"65.3"},"uses_additional_data":false,"paper_date":"2023-11-24","paper":"/paper/oneformer3d-one-transformer-for-unified-point","paper_url":"https://arxiv.org/abs/2311.14405v1","paper_title":"OneFormer3D: One Transformer for Unified Point Cloud Segmentation","code":"https://github.com/oneformer3d/oneformer3d","n_code_links":1,"syntology":null},{"rank_in_archive_order":6,"model":"BFL","metrics":{"mAP@0.25":"74.6","mAP@0.5":"65.3"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":7,"model":"Swin3D-L+CAGroup3D","metrics":{"mAP@0.25":"76.4","mAP@0.5":"63.2"},"uses_additional_data":true,"paper_date":"2023-04-14","paper":"/paper/swin3d-a-pretrained-transformer-backbone-for","paper_url":"https://arxiv.org/abs/2304.06906v3","paper_title":"Swin3D: A Pretrained Transformer Backbone for 3D Indoor Scene Understanding","code":"https://github.com/Pointcept/Pointcept","n_code_links":2,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":8,"model":"QueryFormer","metrics":{"mAP@0.25":"73.4","mAP@0.5":"61.7"},"uses_additional_data":false,"paper_date":"2023-01-01","paper":"/paper/query-refinement-transformer-for-3d-instance","paper_url":"http://openaccess.thecvf.com//content/ICCV2023/html/Lu_Query_Refinement_Transformer_for_3D_Instance_Segmentation_ICCV_2023_paper.html","paper_title":"Query Refinement Transformer for 3D Instance Segmentation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":9,"model":"CAGroup3D","metrics":{"mAP@0.25":"75.1","mAP@0.5":"61.3"},"uses_additional_data":false,"paper_date":"2022-10-09","paper":"/paper/cagroup3d-class-aware-grouping-for-3d-object","paper_url":"https://arxiv.org/abs/2210.04264v1","paper_title":"CAGroup3D: Class-Aware Grouping for 3D Object Detection on Point Clouds","code":"https://github.com/haiyang-w/cagroup3d","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"PBNet","metrics":{"mAP@0.25":"69.3","mAP@0.5":"60.1"},"uses_additional_data":false,"paper_date":"2022-07-22","paper":"/paper/divide-and-conquer-3d-point-cloud-instance","paper_url":"https://arxiv.org/abs/2207.11209v4","paper_title":"Divide and Conquer: 3D Point Cloud Instance Segmentation With Point-Wise Binarization","code":"https://github.com/weiguangzhao/PBNet","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":2,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":11,"model":"SPGroup3D","metrics":{"mAP@0.25":"74.3","mAP@0.5":"59.6"},"uses_additional_data":false,"paper_date":"2023-12-21","paper":"/paper/spgroup3d-superpoint-grouping-network-for","paper_url":"https://arxiv.org/abs/2312.13641v1","paper_title":"SPGroup3D: Superpoint Grouping Network for Indoor 3D Object Detection","code":"https://github.com/zyrant/spgroup3d","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"Point-GCC+TR3D","metrics":{"mAP@0.25":"73.1","mAP@0.5":"59.6"},"uses_additional_data":false,"paper_date":"2023-05-31","paper":"/paper/point-gcc-universal-self-supervised-3d-scene","paper_url":"https://arxiv.org/abs/2305.19623v2","paper_title":"Point-GCC: Universal Self-supervised 3D Scene Pre-training via Geometry-Color Contrast","code":"https://github.com/asterisci/point-gcc","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"SoftGroup","metrics":{"mAP@0.25":"71.6","mAP@0.5":"59.4"},"uses_additional_data":false,"paper_date":"2022-03-03","paper":"/paper/softgroup-for-3d-instance-segmentation-on","paper_url":"https://arxiv.org/abs/2203.01509v1","paper_title":"SoftGroup for 3D Instance Segmentation on Point Clouds","code":"https://github.com/thangvubk/softgroup","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"TR3D","metrics":{"mAP@0.25":"72.9","mAP@0.5":"59.3"},"uses_additional_data":false,"paper_date":"2023-02-06","paper":"/paper/tr3d-towards-real-time-indoor-3d-object","paper_url":"https://arxiv.org/abs/2302.02858v3","paper_title":"TR3D: Towards Real-Time Indoor 3D Object Detection","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":15,"model":"DEST (based on  GroupFree3D)","metrics":{"mAP@0.25":"71.3","mAP@0.5":"58.1"},"uses_additional_data":false,"paper_date":"2025-03-18","paper":"/paper/state-space-model-meets-transformer-a-new-1","paper_url":"https://arxiv.org/abs/2503.14493v1","paper_title":"State Space Model Meets Transformer: A New Paradigm for 3D Object Detection","code":"https://github.com/OpenSpaceAI/DEST3D","n_code_links":1,"syntology":{"n_ran":6,"n_unverified":4,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":16,"model":"FCAF3D","metrics":{"mAP@0.25":"71.5","mAP@0.5":"57.3"},"uses_additional_data":false,"paper_date":"2021-12-01","paper":"/paper/fcaf3d-fully-convolutional-anchor-free-3d","paper_url":"https://arxiv.org/abs/2112.00322v2","paper_title":"FCAF3D: Fully Convolutional Anchor-Free 3D Object Detection","code":"https://github.com/open-mmlab/mmdetection3d","n_code_links":2,"syntology":null},{"rank_in_archive_order":17,"model":"RBGNet","metrics":{"mAP@0.25":"70.6","mAP@0.5":"55.2"},"uses_additional_data":false,"paper_date":"2022-04-05","paper":"/paper/rbgnet-ray-based-grouping-for-3d-object","paper_url":"https://arxiv.org/abs/2204.02251v1","paper_title":"RBGNet: Ray-based Grouping for 3D Object Detection","code":"https://github.com/haiyang-w/rbgnet","n_code_links":1,"syntology":null},{"rank_in_archive_order":18,"model":"RepSurf-U","metrics":{"mAP@0.25":"71.2","mAP@0.5":"54.8"},"uses_additional_data":false,"paper_date":"2022-05-11","paper":"/paper/surface-representation-for-point-clouds","paper_url":"https://arxiv.org/abs/2205.05740v2","paper_title":"Surface Representation for Point Clouds","code":"https://github.com/hancyran/RepSurf","n_code_links":1,"syntology":{"n_ran":13,"n_unverified":2,"n_samples":15,"n_pointer_only_licence":0}},{"rank_in_archive_order":19,"model":"TokenFusion","metrics":{"mAP@0.25":"70.8","mAP@0.5":"54.2"},"uses_additional_data":false,"paper_date":"2022-04-19","paper":"/paper/multimodal-token-fusion-for-vision","paper_url":"https://arxiv.org/abs/2204.08721v2","paper_title":"Multimodal Token Fusion for Vision Transformers","code":"https://github.com/huawei-noah/noah-research/tree/master/TokenFusion","n_code_links":11,"syntology":{"n_ran":1,"n_unverified":2,"n_samples":3,"n_pointer_only_licence":1}},{"rank_in_archive_order":20,"model":"GroupFree3D","metrics":{"mAP@0.25":"69.1","mAP@0.5":"52.8"},"uses_additional_data":false,"paper_date":"2021-04-01","paper":"/paper/group-free-3d-object-detection-via","paper_url":"https://arxiv.org/abs/2104.00678v2","paper_title":"Group-Free 3D Object Detection via Transformers","code":"https://github.com/zeliu98/Group-Free-3D","n_code_links":4,"syntology":{"n_ran":9,"n_unverified":8,"n_samples":17,"n_pointer_only_licence":0}},{"rank_in_archive_order":21,"model":"BRNet","metrics":{"mAP@0.25":"66.1","mAP@0.5":"50.9"},"uses_additional_data":false,"paper_date":"2021-04-13","paper":"/paper/back-tracing-representative-points-for-voting","paper_url":"https://arxiv.org/abs/2104.06114v2","paper_title":"Back-tracing Representative Points for Voting-based 3D Object Detection in Point Clouds","code":"https://github.com/cheng052/BRNet","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":2,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":22,"model":"3D-MPA","metrics":{"mAP@0.25":"64.2","mAP@0.5":"49.2"},"uses_additional_data":false,"paper_date":"2020-03-30","paper":"/paper/3d-mpa-multi-proposal-aggregation-for-3d","paper_url":"https://arxiv.org/abs/2003.13867v1","paper_title":"3D-MPA: Multi Proposal Aggregation for 3D Semantic Instance Segmentation","code":"https://github.com/francisengelmann/3D-MPA","n_code_links":1,"syntology":null},{"rank_in_archive_order":23,"model":"H3DNet","metrics":{"mAP@0.25":"67.2","mAP@0.5":"48.1"},"uses_additional_data":false,"paper_date":"2020-06-10","paper":"/paper/h3dnet-3d-object-detection-using-hybrid","paper_url":"https://arxiv.org/abs/2006.05682v3","paper_title":"H3DNet: 3D Object Detection Using Hybrid Geometric Primitives","code":"https://github.com/open-mmlab/mmdetection3d","n_code_links":2,"syntology":null},{"rank_in_archive_order":24,"model":"3DETR-m","metrics":{"mAP@0.25":"65.0","mAP@0.5":"47.0"},"uses_additional_data":false,"paper_date":"2021-09-16","paper":"/paper/an-end-to-end-transformer-model-for-3d-object","paper_url":"https://arxiv.org/abs/2109.08141v1","paper_title":"An End-to-End Transformer Model for 3D Object Detection","code":"https://github.com/facebookresearch/3detr","n_code_links":1,"syntology":{"n_ran":6,"n_unverified":1,"n_samples":7,"n_pointer_only_licence":1}},{"rank_in_archive_order":25,"model":"VoteNet (PC-FractalDB)","metrics":{"mAP@0.25":"63.4","mAP@0.5":"39.9"},"uses_additional_data":true,"paper_date":"2022-01-01","paper":"/paper/point-cloud-pre-training-with-natural-3d","paper_url":"http://openaccess.thecvf.com//content/CVPR2022/html/Yamada_Point_Cloud_Pre-Training_With_Natural_3D_Structures_CVPR_2022_paper.html","paper_title":"Point Cloud Pre-Training With Natural 3D Structures","code":"https://github.com/ryosuke-yamada/3dfractaldb","n_code_links":1,"syntology":null},{"rank_in_archive_order":26,"model":"GSDN","metrics":{"mAP@0.25":"62.8","mAP@0.5":"34.8"},"uses_additional_data":false,"paper_date":"2020-06-22","paper":"/paper/generative-sparse-detection-networks-for-3d","paper_url":"https://arxiv.org/abs/2006.12356v1","paper_title":"Generative Sparse Detection Networks for 3D Single-shot Object Detection","code":"https://github.com/StanfordVL/MinkowskiEngine","n_code_links":4,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":27,"model":"HGNet","metrics":{"mAP@0.25":"61.3","mAP@0.5":"34.4"},"uses_additional_data":false,"paper_date":"2020-06-01","paper":"/paper/a-hierarchical-graph-network-for-3d-object","paper_url":"http://openaccess.thecvf.com/content_CVPR_2020/html/Chen_A_Hierarchical_Graph_Network_for_3D_Object_Detection_on_Point_CVPR_2020_paper.html","paper_title":"A Hierarchical Graph Network for 3D Object Detection on Point Clouds","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":28,"model":"VoteNet","metrics":{"mAP@0.25":"58.6","mAP@0.5":"33.5"},"uses_additional_data":false,"paper_date":"2019-04-21","paper":"/paper/deep-hough-voting-for-3d-object-detection-in","paper_url":"https://arxiv.org/abs/1904.09664v2","paper_title":"Deep Hough Voting for 3D Object Detection in Point Clouds","code":"https://github.com/open-mmlab/mmdetection3d","n_code_links":13,"syntology":{"n_ran":5,"n_unverified":5,"n_samples":10,"n_pointer_only_licence":10}},{"rank_in_archive_order":29,"model":"ImGeoNet (RGB only)","metrics":{"mAP@0.25":"54.8","mAP@0.5":"28.4"},"uses_additional_data":false,"paper_date":"2023-08-17","paper":"/paper/imgeonet-image-induced-geometry-aware-voxel","paper_url":"https://arxiv.org/abs/2308.09098v1","paper_title":"ImGeoNet: Image-induced Geometry-aware Voxel Representation for Multi-view 3D Object Detection","code":"https://github.com/ttaoREtw/ImGeoNet","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":30,"model":"ImVoxelNet (RGB only)","metrics":{"mAP@0.25":"48.1","mAP@0.5":"22.7"},"uses_additional_data":false,"paper_date":"2021-06-02","paper":"/paper/imvoxelnet-image-to-voxels-projection-for","paper_url":"https://arxiv.org/abs/2106.01178v3","paper_title":"ImVoxelNet: Image to Voxels Projection for Monocular and Multi-View General-Purpose 3D Object Detection","code":"https://github.com/open-mmlab/mmdetection3d","n_code_links":2,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":31,"model":"3D-SIS","metrics":{"mAP@0.25":"40.2","mAP@0.5":"22.5"},"uses_additional_data":false,"paper_date":"2018-12-17","paper":"/paper/3d-sis-3d-semantic-instance-segmentation-of","paper_url":"http://arxiv.org/abs/1812.07003v3","paper_title":"3D-SIS: 3D Semantic Instance Segmentation of RGB-D Scans","code":"https://github.com/Sekunde/3D-SIS","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":2,"n_samples":2,"n_pointer_only_licence":2}},{"rank_in_archive_order":32,"model":"GSPN","metrics":{"mAP@0.25":"30.6","mAP@0.5":"17.7"},"uses_additional_data":false,"paper_date":"2018-12-08","paper":"/paper/gspn-generative-shape-proposal-network-for-3d","paper_url":"http://arxiv.org/abs/1812.03320v1","paper_title":"GSPN: Generative Shape Proposal Network for 3D Instance Segmentation in Point Cloud","code":"https://github.com/ericyi/GSPN","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":8,"n_samples":8,"n_pointer_only_licence":0}},{"rank_in_archive_order":33,"model":"SGPN","metrics":{"mAP@0.25":"20.7"},"uses_additional_data":false,"paper_date":"2017-11-23","paper":"/paper/sgpn-similarity-group-proposal-network-for-3d","paper_url":"https://arxiv.org/abs/1711.08588v2","paper_title":"SGPN: Similarity Group Proposal Network for 3D Point Cloud Instance Segmentation","code":"https://github.com/laughtervv/SGPN","n_code_links":1,"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 6,885 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":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"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":16,"rows_with_any_sample_ran":14,"distinct_papers_with_graph_line":15,"distinct_papers_with_any_sample_ran":13,"samples_over_distinct_papers":{"n_ran":63,"n_unverified":40,"n_samples":103,"n_pointer_only_licence":20,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":69,"n_unverified":44,"n_samples":113,"n_pointer_only_licence":20,"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"}}}