{"url":"/dataset/h2o-dataset","name":"H2O  (2 Hands and Objects)","full_name":null,"description_markdown":"We present a comprehensive framework for egocentric interaction recognition using markerless 3D annotations of two hands manipulating objects. To this end, we propose a method to create a unified dataset for egocentric 3D interaction recognition. Our method produces annotations of the 3D pose of two hands and the 6D pose of the manipulated objects, along with their interaction labels for each frame. Our dataset, called H2O (2 Hands and Objects), provides synchronized multi-view RGB-D images, interaction labels, object classes, ground-truth 3D poses for left & right hands, 6D object poses, ground-truth camera poses, object meshes and scene point clouds. To the best of our knowledge, this is the first benchmark that enables the study of first-person actions with the use of the pose of both left and right hands manipulating objects and presents an unprecedented level of detail for egocentric 3D interaction recognition. We further propose the method to predict interaction classes by estimating the 3D pose of two hands and the 6D pose of the manipulated objects, jointly from RGB images. Our method models both inter- and intra-dependencies between both hands and objects by learning the topology of a graph convolutional network that predicts interactions. We show that our method facilitated by this dataset establishes a strong baseline for joint hand-object pose estimation and achieves state-of-the-art accuracy for first person interaction recognition.","description_withheld":null,"homepage":"https://taeinkwon.com/projects/h2o/","introduced_date":"2021-04-21","introduced_date_note":null,"introduced_by":null,"license":null,"modalities":[],"tasks":[{"name":"Action Recognition","url":"/task/action-recognition-in-videos","datasets_with_task":"/datasets/task/action-recognition-in-videos"},{"name":"Skeleton Based Action Recognition","url":"/task/skeleton-based-action-recognition","datasets_with_task":"/datasets/task/skeleton-based-action-recognition"},{"name":"3D Hand Pose Estimation","url":"/task/3d-hand-pose-estimation","datasets_with_task":"/datasets/task/3d-hand-pose-estimation"}],"languages":[],"variants":["H2O  (2 Hands and Objects)"],"data_loaders":[],"num_papers_in_archive":14,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/action-recognition-on-h2o-2-hands-and-objects","task":"Action Recognition","dataset_variant":"H2O  (2 Hands and Objects)","rows":11,"metrics":["Actions Top-1","RGB","Hand Pose","Object Pose","Object Label"],"first_row_in_archive_order":{"model":"HandFormer-B/21x8","paper":"/paper/on-the-utility-of-3d-hand-poses-for-action","metrics":{"Actions Top-1":"93.39","Hand Pose":"3D","Object Label":"No","Object Pose":"No","RGB":"Yes"},"code_links":[{"title":"s-shamil/HandFormer","url":"https://github.com/s-shamil/HandFormer"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/skeleton-based-action-recognition-on-h2o-2","task":"Skeleton Based Action Recognition","dataset_variant":"H2O  (2 Hands and Objects)","rows":4,"metrics":["Accuracy"],"first_row_in_archive_order":{"model":"CHASE(STSA-Net)","paper":"/paper/chase-learning-convex-hull-adaptive-shift-for","metrics":{"Accuracy":"94.77"},"code_links":[{"title":"Necolizer/CHASE","url":"https://github.com/Necolizer/CHASE"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/chase-learning-convex-hull-adaptive-shift-for","title":"CHASE: Learning Convex Hull Adaptive Shift for Skeleton-based Multi-Entity Action Recognition","date":"2024-10-09","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":2,"samples_ran":2,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/sharp-segmentation-of-hands-and-arms-by-range","title":"SHARP: Segmentation of Hands and Arms by Range using Pseudo-Depth for Enhanced Egocentric 3D Hand Pose Estimation and Action Recognition","date":"2024-08-19","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/in-my-perspective-in-my-hands-accurate","title":"In My Perspective, In My Hands: Accurate Egocentric 2D Hand Pose and Action Recognition","date":"2024-04-14","rows_on_this_dataset":2,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":5,"samples_ran":4,"samples_unverified":1,"pointer_only_for_licence":5,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/on-the-utility-of-3d-hand-poses-for-action","title":"On the Utility of 3D Hand Poses for Action Recognition","date":"2024-03-14","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":13,"samples_ran":8,"samples_unverified":5,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/interactive-spatiotemporal-token-attention","title":"Interactive Spatiotemporal Token Attention Network for Skeleton-based General Interactive Action Recognition","date":"2023-07-14","rows_on_this_dataset":2,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":1,"samples_ran":1,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/transformer-based-unified-recognition-of-two","title":"Transformer-Based Unified Recognition of Two Hands Manipulating Objects","date":"2023-01-01","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/hierarchical-temporal-transformer-for-3d-hand","title":"Hierarchical Temporal Transformer for 3D Hand Pose Estimation and Action Recognition from Egocentric RGB Videos","date":"2022-09-20","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":8,"samples_ran":1,"samples_unverified":7,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/revisiting-skeleton-based-action-recognition","title":"Revisiting Skeleton-based Action Recognition","date":"2021-04-28","rows_on_this_dataset":1,"code_links":4,"syntology":null},{"paper":"/paper/h2o-two-hands-manipulating-objects-for-first","title":"H2O: Two Hands Manipulating Objects for First Person Interaction Recognition","date":"2021-04-22","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/disentangling-and-unifying-graph-convolutions","title":"Disentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognition","date":"2020-03-31","rows_on_this_dataset":1,"code_links":3,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":3,"samples_ran":3,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/slowfast-networks-for-video-recognition","title":"SlowFast Networks for Video Recognition","date":"2018-12-10","rows_on_this_dataset":1,"code_links":15,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":10,"samples_ran":0,"samples_unverified":10,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/spatial-temporal-graph-convolutional-networks-1","title":"Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition","date":"2018-01-23","rows_on_this_dataset":1,"code_links":24,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":2,"samples_ran":2,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":8,"samples_harvested":44,"samples_ran":21,"samples_unverified":23,"pointer_only_for_licence":5,"papers_with_no_sample_that_ran":1,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}