Large-scale manipulation data is essential for robot learning, yet collecting robot demonstrations remains expensive and difficult to scale. Meanwhile, abundant egocentric human manipulation videos provide rich behavioral experiences, but transferring them across embodiments remains challenging due to differences between human hands and robotic end-effectors. Recent advances in video world models offer a promising pathway to synthesize robot-centric manipulation videos from human observations, while their cross-embodiment transfer capability remains largely unexplored. Therefore, we introduce H2R-Bench, a benchmark for evaluating cross-embodiment human-to-robot manipulation video generation, where models transform egocentric human demonstrations into robot manipulation videos under specified embodiments. Each benchmark instance contains a human demonstration video, target embodiment constraints, and source-grounded annotations covering task goals, action events, functional contacts, and object responses. H2R-Bench evaluates generated videos through five dimensions, including goal-state completion, action-event completion, functional contact transfer, embodiment correctness, and general video quality. We benchmark eleven stateof-the-art video generation models across six manipulation families and two robot embodiments. Our evaluation reveals that current video world models remain limited in human-torobot manipulation transfer: even leading models often fail in embodiment consistency, functional interaction, and task execution. H2R-Bench provides a systematic diagnostic framework for evaluating whether video world models can bridge the human-to-robot embodiment gap and convert human manipulation observations into robot-centric training resources.</p>\n<p><a href=\"https://cdn-uploads.huggingface.co/production/uploads/63048965eb6d777a838cb7a8/d32q0jPQNf8mApagcDJ_N.png\" rel=\"nofollow\"><img src=\"https://cdn-uploads.huggingface.co/production/uploads/63048965eb6d777a838cb7a8/d32q0jPQNf8mApagcDJ_N.png\" alt=\"Clipboard_Screenshot_1786680870\"></a></p>\n<p><a href=\"https://cdn-uploads.huggingface.co/production/uploads/63048965eb6d777a838cb7a8/eMSQrbjhHLXiKjLGW7C0e.png\" rel=\"nofollow\"><img src=\"https://cdn-uploads.huggingface.co/production/uploads/63048965eb6d777a838cb7a8/eMSQrbjhHLXiKjLGW7C0e.png\" alt=\"Clipboard_Screenshot_1786680893\"></a></p>\n","updatedAt":"2026-08-14T04:16:21.468Z","author":{"_id":"63048965eb6d777a838cb7a8","avatarUrl":"/avatars/b987fb7f630443bf94a03daf8dcbffe9.svg","fullname":"chaofanma","name":"chaofanma","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8105422258377075},"editors":["chaofanma"],"editorAvatarUrls":["/avatars/b987fb7f630443bf94a03daf8dcbffe9.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.13049","authors":[{"_id":"6a7e93ae42823931a1f176cb","name":"Dingyi Rong","hidden":false},{"_id":"6a7e93ae42823931a1f176cc","name":"Yue Shi","hidden":false},{"_id":"6a7e93ae42823931a1f176cd","name":"Chaofan Ma","hidden":false},{"_id":"6a7e93ae42823931a1f176ce","name":"Jiezhang Cao","hidden":false},{"_id":"6a7e93ae42823931a1f176cf","name":"Zongrui Wang","hidden":false},{"_id":"6a7e93ae42823931a1f176d0","name":"Zeyu Zhang","hidden":false},{"_id":"6a7e93ae42823931a1f176d1","name":"Yao Mu","hidden":false},{"_id":"6a7e93ae42823931a1f176d2","name":"Guangtao Zhai","hidden":false},{"_id":"6a7e93ae42823931a1f176d3","name":"Ning Liu","hidden":false}],"publishedAt":"2026-08-13T00:00:00.000Z","submittedOnDailyAt":"2026-08-14T00:00:00.000Z","title":"H2R-Bench: Benchmarking Human-to-Robot Manipulation Video Generation in World Models","submittedOnDailyBy":{"_id":"63048965eb6d777a838cb7a8","avatarUrl":"/avatars/b987fb7f630443bf94a03daf8dcbffe9.svg","isPro":false,"fullname":"chaofanma","user":"chaofanma","type":"user","name":"chaofanma"},"summary":"Large-scale manipulation data is essential for robot learning, yet collecting robot demonstrations remains expensive and difficult to scale. Meanwhile, abundant egocentric human manipulation videos provide rich behavioral experiences, but transferring them across embodiments remains challenging due to differences between human hands and robotic end-effectors. Recent advances in video world models offer a promising pathway to synthesize robot-centric manipulation videos from human observations, while their cross-embodiment transfer capability remains largely unexplored. Therefore, we introduce H2R-Bench, a benchmark for evaluating cross-embodiment human-to-robot manipulation video generation, where models transform egocentric human demonstrations into robot manipulation videos under specified embodiments. Each benchmark instance contains a human demonstration video, target embodiment constraints, and source-grounded annotations covering task goals, action events, functional contacts, and object responses. H2R-Bench evaluates generated videos through five dimensions, including goal-state completion, action-event completion, functional contact transfer, embodiment correctness, and general video quality. We benchmark eleven state-of-the-art video generation models across six manipulation families and two robot embodiments. Our evaluation reveals that current video world models remain limited in human-to-robot manipulation transfer: even leading models often fail in embodiment consistency, functional interaction, and task execution. H2R-Bench provides a systematic diagnostic framework for evaluating whether video world models can bridge the human-to-robot embodiment gap and convert human manipulation observations into robot-centric training resources.","upvotes":5,"discussionId":"6a7e93af42823931a1f176d4","projectPage":"https://rongdingyi.github.io/H2R-Bench/","githubRepo":"https://github.com/Rongdingyi/H2R-Bench","githubRepoAddedBy":"user","ai_summary":"H2R-Bench evaluates video generation models on transforming human manipulation videos into robot-centric demonstrations across embodiment constraints and interaction fidelity.","ai_keywords":["video world models","cross-embodiment transfer","human-to-robot manipulation","egocentric video","embodiment constraints","functional contact","action events","goal-state completion"],"ai_summary_model":"thinkingmachines/Inkling-Small","githubStars":0,"organization":{"_id":"63e5ef7bf2e9a8f22c515654","name":"SJTU","fullname":"Shanghai Jiao Tong University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1676013394657-63e5ee22b6a40bf941da0928.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"63048965eb6d777a838cb7a8","avatarUrl":"/avatars/b987fb7f630443bf94a03daf8dcbffe9.svg","isPro":false,"fullname":"chaofanma","user":"chaofanma","type":"user"},{"_id":"675a621105c46a17e6a229b3","avatarUrl":"/avatars/1b51f141df9d206fcbd2598b6e994aa6.svg","isPro":false,"fullname":"Dingyi Rong","user":"dingyi11","type":"user"},{"_id":"6a069177a01745697eb21189","avatarUrl":"/avatars/2cff7b99d89036a327c47e1cf3220617.svg","isPro":false,"fullname":"shiyue001","user":"shiyue0011","type":"user"},{"_id":"6655d5575b8ab1ed4f66265d","avatarUrl":"/avatars/1fd6da28eba1c804cad1cc490b374eac.svg","isPro":true,"fullname":"Chen Ye","user":"sjtuchenye","type":"user"},{"_id":"6731af65389aca4be7ce8a75","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/6Ym2bfkiJzKOtDZ3LCdFg.png","isPro":false,"fullname":"Cumulus","user":"CumulusAlpha","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"63e5ef7bf2e9a8f22c515654","name":"SJTU","fullname":"Shanghai Jiao Tong University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1676013394657-63e5ee22b6a40bf941da0928.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.13049.md","query":{}}">
H2R-Bench: Benchmarking Human-to-Robot Manipulation Video Generation in World Models
Abstract
H2R-Bench evaluates video generation models on transforming human manipulation videos into robot-centric demonstrations across embodiment constraints and interaction fidelity.
Large-scale manipulation data is essential for robot learning, yet collecting robot demonstrations remains expensive and difficult to scale. Meanwhile, abundant egocentric human manipulation videos provide rich behavioral experiences, but transferring them across embodiments remains challenging due to differences between human hands and robotic end-effectors. Recent advances in video world models offer a promising pathway to synthesize robot-centric manipulation videos from human observations, while their cross-embodiment transfer capability remains largely unexplored. Therefore, we introduce H2R-Bench, a benchmark for evaluating cross-embodiment human-to-robot manipulation video generation, where models transform egocentric human demonstrations into robot manipulation videos under specified embodiments. Each benchmark instance contains a human demonstration video, target embodiment constraints, and source-grounded annotations covering task goals, action events, functional contacts, and object responses. H2R-Bench evaluates generated videos through five dimensions, including goal-state completion, action-event completion, functional contact transfer, embodiment correctness, and general video quality. We benchmark eleven state-of-the-art video generation models across six manipulation families and two robot embodiments. Our evaluation reveals that current video world models remain limited in human-to-robot manipulation transfer: even leading models often fail in embodiment consistency, functional interaction, and task execution. H2R-Bench provides a systematic diagnostic framework for evaluating whether video world models can bridge the human-to-robot embodiment gap and convert human manipulation observations into robot-centric training resources.
Community
Large-scale manipulation data is essential for robot learning, yet collecting robot demonstrations remains expensive and difficult to scale. Meanwhile, abundant egocentric human manipulation videos provide rich behavioral experiences, but transferring them across embodiments remains challenging due to differences between human hands and robotic end-effectors. Recent advances in video world models offer a promising pathway to synthesize robot-centric manipulation videos from human observations, while their cross-embodiment transfer capability remains largely unexplored. Therefore, we introduce H2R-Bench, a benchmark for evaluating cross-embodiment human-to-robot manipulation video generation, where models transform egocentric human demonstrations into robot manipulation videos under specified embodiments. Each benchmark instance contains a human demonstration video, target embodiment constraints, and source-grounded annotations covering task goals, action events, functional contacts, and object responses. H2R-Bench evaluates generated videos through five dimensions, including goal-state completion, action-event completion, functional contact transfer, embodiment correctness, and general video quality. We benchmark eleven stateof-the-art video generation models across six manipulation families and two robot embodiments. Our evaluation reveals that current video world models remain limited in human-torobot manipulation transfer: even leading models often fail in embodiment consistency, functional interaction, and task execution. H2R-Bench provides a systematic diagnostic framework for evaluating whether video world models can bridge the human-to-robot embodiment gap and convert human manipulation observations into robot-centric training resources.
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2608.13049 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.13049 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.13049 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.