https://huggingface.co/tencent/Hunyuan3D-Part

\n","updatedAt":"2025-09-26T13:08:31.813Z","author":{"_id":"63a369d98c0c89dcae3b8329","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63a369d98c0c89dcae3b8329/AiH2zjy1cnt9OADAAZMLD.jpeg","fullname":"Adina Yakefu","name":"AdinaY","type":"user","isPro":true,"isHf":true,"isHfAdmin":false,"isMod":false,"followerCount":1269}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.738471508026123},"editors":["AdinaY"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/63a369d98c0c89dcae3b8329/AiH2zjy1cnt9OADAAZMLD.jpeg"],"reactions":[],"isReport":false,"parentCommentId":"68c2359a512125ad6b9de466"}},{"id":"68d6b1b96667bc27ccbf9272","author":{"_id":"68d6af7244856bdfd842aa2a","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/oD0cx6ZEtZOm2NyiBAHXH.png","fullname":"sai kshipra budihal","name":"Saikshipra","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false},"createdAt":"2025-09-26T15:31:05.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"Impressive work. Automating robust 3D part segmentation at this scale is a huge leap for design, robotics, and digital twins. Excited to see how P3-SAM accelerates real-world 3D workflows once the code is out.","html":"

Impressive work. Automating robust 3D part segmentation at this scale is a huge leap for design, robotics, and digital twins. Excited to see how P3-SAM accelerates real-world 3D workflows once the code is out.

\n","updatedAt":"2025-09-26T15:31:05.209Z","author":{"_id":"68d6af7244856bdfd842aa2a","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/oD0cx6ZEtZOm2NyiBAHXH.png","fullname":"sai kshipra budihal","name":"Saikshipra","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.9249698519706726},"editors":["Saikshipra"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/oD0cx6ZEtZOm2NyiBAHXH.png"],"reactions":[],"isReport":false,"parentCommentId":"68c2359a512125ad6b9de466"}}]},{"id":"68c378a7b916caa8b5ffd5a6","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":264},"createdAt":"2025-09-12T01:34:31.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [GeoSAM2: Unleashing the Power of SAM2 for 3D Part Segmentation](https://huggingface.co/papers/2508.14036) (2025)\n* [SGS-3D: High-Fidelity 3D Instance Segmentation via Reliable Semantic Mask Splitting and Growing](https://huggingface.co/papers/2509.05144) (2025)\n* [Integrating SAM Supervision for 3D Weakly Supervised Point Cloud Segmentation](https://huggingface.co/papers/2508.19909) (2025)\n* [Details Matter for Indoor Open-vocabulary 3D Instance Segmentation](https://huggingface.co/papers/2507.23134) (2025)\n* [X-Part: high fidelity and structure coherent shape decomposition](https://huggingface.co/papers/2509.08643) (2025)\n* [PointGauss: Point Cloud-Guided Multi-Object Segmentation for Gaussian Splatting](https://huggingface.co/papers/2508.00259) (2025)\n* [SAGOnline: Segment Any Gaussians Online](https://huggingface.co/papers/2508.08219) (2025)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"

This is an automated message from the Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2025-09-12T01:34:31.890Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":264}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7015725374221802},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}},{"id":"68d65de8f854937280854747","author":{"_id":"66fe2aadd5ede3018ab704bf","avatarUrl":"/avatars/46f2ae3717d1dfa5773d419e4dc9a891.svg","fullname":"Changhangfeng MA","name":"murcherful","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false},"createdAt":"2025-09-26T09:33:28.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"Our code and demo are released!!","html":"

Our code and demo are released!!

\n","updatedAt":"2025-09-26T09:33:28.431Z","author":{"_id":"66fe2aadd5ede3018ab704bf","avatarUrl":"/avatars/46f2ae3717d1dfa5773d419e4dc9a891.svg","fullname":"Changhangfeng MA","name":"murcherful","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.9730382561683655},"editors":["murcherful"],"editorAvatarUrls":["/avatars/46f2ae3717d1dfa5773d419e4dc9a891.svg"],"reactions":[{"reaction":"🤗","users":["murcherful"],"count":1}],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2509.06784","authors":[{"_id":"68bfa444207285de11b07c13","user":{"_id":"66fe2aadd5ede3018ab704bf","avatarUrl":"/avatars/46f2ae3717d1dfa5773d419e4dc9a891.svg","isPro":false,"fullname":"Changhangfeng MA","user":"murcherful","type":"user"},"name":"Changfeng Ma","status":"claimed_verified","statusLastChangedAt":"2025-09-09T13:47:28.483Z","hidden":false},{"_id":"68bfa444207285de11b07c14","name":"Yang Li","hidden":false},{"_id":"68bfa444207285de11b07c15","user":{"_id":"6434f9bfdf32a2296635f88d","avatarUrl":"/avatars/dc19ba1080b0b17a220d7e52bd514f13.svg","isPro":false,"fullname":"Xinhao Yan","user":"HowieYan","type":"user"},"name":"Xinhao Yan","status":"claimed_verified","statusLastChangedAt":"2025-09-12T16:20:00.639Z","hidden":false},{"_id":"68bfa444207285de11b07c16","name":"Jiachen Xu","hidden":false},{"_id":"68bfa444207285de11b07c17","name":"Yunhan Yang","hidden":false},{"_id":"68bfa444207285de11b07c18","name":"Chunshi Wang","hidden":false},{"_id":"68bfa444207285de11b07c19","name":"Zibo Zhao","hidden":false},{"_id":"68bfa444207285de11b07c1a","name":"Yanwen Guo","hidden":false},{"_id":"68bfa444207285de11b07c1b","name":"Zhuo Chen","hidden":false},{"_id":"68bfa444207285de11b07c1c","name":"Chunchao Guo","hidden":false}],"publishedAt":"2025-09-08T15:12:17.000Z","submittedOnDailyAt":"2025-09-11T01:06:10.468Z","title":"P3-SAM: Native 3D Part Segmentation","submittedOnDailyBy":{"_id":"66fe2aadd5ede3018ab704bf","avatarUrl":"/avatars/46f2ae3717d1dfa5773d419e4dc9a891.svg","isPro":false,"fullname":"Changhangfeng MA","user":"murcherful","type":"user"},"summary":"Segmenting 3D assets into their constituent parts is crucial for enhancing 3D\nunderstanding, facilitating model reuse, and supporting various applications\nsuch as part generation. However, current methods face limitations such as poor\nrobustness when dealing with complex objects and cannot fully automate the\nprocess. In this paper, we propose a native 3D point-promptable part\nsegmentation model termed P3-SAM, designed to fully automate the segmentation\nof any 3D objects into components. Inspired by SAM, P3-SAM consists of a\nfeature extractor, multiple segmentation heads, and an IoU predictor, enabling\ninteractive segmentation for users. We also propose an algorithm to\nautomatically select and merge masks predicted by our model for part instance\nsegmentation. Our model is trained on a newly built dataset containing nearly\n3.7 million models with reasonable segmentation labels. Comparisons show that\nour method achieves precise segmentation results and strong robustness on any\ncomplex objects, attaining state-of-the-art performance. Our code will be\nreleased soon.","upvotes":21,"discussionId":"68bfa444207285de11b07c1d","projectPage":"https://murcherful.github.io/P3-SAM/","githubRepo":"https://github.com/Tencent-Hunyuan/Hunyuan3D-Part/tree/main/P3-SAM","ai_summary":"P3-SAM, a native 3D point-promptable part segmentation model, achieves precise and robust segmentation of complex 3D objects using a feature extractor, multiple segmentation heads, and an IoU predictor.","ai_keywords":["3D point-promptable part segmentation","P3-SAM","feature extractor","segmentation heads","IoU predictor","interactive segmentation","part instance segmentation","dataset"],"githubStars":195},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"65c1d4f04f21259d9944593c","avatarUrl":"/avatars/355d84485da1032df3a692de6549fb7c.svg","isPro":false,"fullname":"yamaguchi","user":"yamaguchi1997","type":"user"},{"_id":"66fe2aadd5ede3018ab704bf","avatarUrl":"/avatars/46f2ae3717d1dfa5773d419e4dc9a891.svg","isPro":false,"fullname":"Changhangfeng MA","user":"murcherful","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"677e2e55c27c05b2d16f302e","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/s2YxCeBRQShgI73hvdXHR.png","isPro":false,"fullname":"Zohn Wu","user":"Hanco9","type":"user"},{"_id":"6343e37f73b4f9cedab1c846","avatarUrl":"/avatars/2638af4626e8a4e3a95f845b94ad94f6.svg","isPro":false,"fullname":"Leheng Li","user":"lilelife","type":"user"},{"_id":"686fc1fb779dd387eb29a46b","avatarUrl":"/avatars/710c34ac3404426dac4cc6221e29dcdd.svg","isPro":false,"fullname":"Mike Green","user":"AiEson2","type":"user"},{"_id":"6434f9bfdf32a2296635f88d","avatarUrl":"/avatars/dc19ba1080b0b17a220d7e52bd514f13.svg","isPro":false,"fullname":"Xinhao Yan","user":"HowieYan","type":"user"},{"_id":"64b4eecf2fc8324fcb63b404","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64b4eecf2fc8324fcb63b404/zGYqYVB4-o-GBMybJ8CDA.png","isPro":false,"fullname":"Yunhan Yang","user":"yhyang-myron","type":"user"},{"_id":"6863a000e7d0e009965a687b","avatarUrl":"/avatars/60e495c0a1ed9a2ede7a23e0209044c6.svg","isPro":false,"fullname":"yan","user":"xiruii","type":"user"},{"_id":"68c2b6f91be5880077735969","avatarUrl":"/avatars/edf23338b3e88547607607c48cf371d1.svg","isPro":false,"fullname":"Jingyu Gong","user":"gongjingyuair","type":"user"},{"_id":"6787a306fe076b7cbc501815","avatarUrl":"/avatars/8dfe7c61511c1ce78d5b2fe119e8562b.svg","isPro":false,"fullname":"js tang","user":"tjs9707","type":"user"},{"_id":"6407e5294edf9f5c4fd32228","avatarUrl":"/avatars/8e2d55460e9fe9c426eb552baf4b2cb0.svg","isPro":false,"fullname":"Stoney Kang","user":"sikang99","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0,"submitterOrganization":{"_id":"6645f953c39288df638dbdd5","name":"Tencent-Hunyuan","fullname":"Tencent Hunyuan","avatar":"https://cdn-uploads.huggingface.co/production/uploads/62d22496c58f969c152bcefd/woKSjt2wXvBNKussyYPsa.png"}}">

Papers

arxiv:2509.06784

P3-SAM: Native 3D Part Segmentation

Published on Sep 8

· Submitted by

Changhangfeng MA on Sep 11

Tencent Hunyuan

Upvote

Authors:

Changfeng Ma ,

Xinhao Yan ,

Abstract

P3-SAM, a native 3D point-promptable part segmentation model, achieves precise and robust segmentation of complex 3D objects using a feature extractor, multiple segmentation heads, and an IoU predictor.

AI-generated summary

Segmenting 3D assets into their constituent parts is crucial for enhancing 3D understanding, facilitating model reuse, and supporting various applications such as part generation. However, current methods face limitations such as poor robustness when dealing with complex objects and cannot fully automate the process. In this paper, we propose a native 3D point-promptable part segmentation model termed P3-SAM, designed to fully automate the segmentation of any 3D objects into components. Inspired by SAM, P3-SAM consists of a feature extractor, multiple segmentation heads, and an IoU predictor, enabling interactive segmentation for users. We also propose an algorithm to automatically select and merge masks predicted by our model for part instance segmentation. Our model is trained on a newly built dataset containing nearly 3.7 million models with reasonable segmentation labels. Comparisons show that our method achieves precise segmentation results and strong robustness on any complex objects, attaining state-of-the-art performance. Our code will be released soon.