Accepted to ECCV 2026.</p>\n","updatedAt":"2026-08-04T06:12:03.031Z","author":{"_id":"66c2f277a2213740f48ff641","avatarUrl":"/avatars/118cd48dc1692db21f281b9a9f13d539.svg","fullname":"changwoo baek","name":"higokri","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":1,"identifiedLanguage":{"language":"en","probability":0.8106326460838318},"editors":["higokri"],"editorAvatarUrls":["/avatars/118cd48dc1692db21f281b9a9f13d539.svg"],"reactions":[],"isReport":false}},{"id":"6a7182a7f28e83d738817404","author":{"_id":"66c2f277a2213740f48ff641","avatarUrl":"/avatars/118cd48dc1692db21f281b9a9f13d539.svg","fullname":"changwoo baek","name":"higokri","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false},"createdAt":"2026-08-04T06:11:51.000Z","type":"comment","data":{"edited":true,"hidden":true,"hiddenBy":"","hiddenReason":"Resolved","latest":{"raw":"This comment has been hidden","html":"This comment has been hidden","updatedAt":"2026-08-04T06:12:45.850Z","author":{"_id":"66c2f277a2213740f48ff641","avatarUrl":"/avatars/118cd48dc1692db21f281b9a9f13d539.svg","fullname":"changwoo baek","name":"higokri","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"editors":[],"editorAvatarUrls":[],"reactions":[]}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.01185","authors":[{"_id":"6a7181abec5082b9f872cf5e","name":"Changwoo Baek","hidden":false},{"_id":"6a7181abec5082b9f872cf5f","name":"Kyeongbo Kong","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/66c2f277a2213740f48ff641/zkhjsdDrrCxjnuhhpFJ6_.mp4"],"publishedAt":"2026-08-02T00:00:00.000Z","submittedOnDailyAt":"2026-08-04T00:00:00.000Z","title":"3DZip: Spatial-Aware Feature Diversity-Guided Token Compression for 3D Question Answering","submittedOnDailyBy":{"_id":"66c2f277a2213740f48ff641","avatarUrl":"/avatars/118cd48dc1692db21f281b9a9f13d539.svg","isPro":false,"fullname":"changwoo baek","user":"higokri","type":"user","name":"higokri"},"summary":"Recent 3D vision-language models (3D VLMs) construct geometry aware tokens by projecting 2D visual features into world coordinates, enabling spatial reasoning for tasks such as 3D question answering. However, this design generates thousands of tokens per scene, resulting in substantial computational and memory overhead. While token compression has been extensively studied in 2D VLMs, existing approaches rely on semantic relevance or attention-based selection that overlook the structured spatial nature of 3D tokens. Moreover, redundancy in 3D representations cannot be resolved by spatial proximity alone, as object-level token imbalance persists even after spatial aggregation. To address this, we propose 3DZip, a three-stage token compression framework that first applies coarse voxelization to remove point-level redundancy, then selects anchor tokens based on feature-space diversity via a Determinantal Point Process, and finally merges remaining tokens under spatial constraints to preserve geometric coherence. Experiments on three 3D question answering benchmarks demonstrate that 3DZip consistently outperforms existing compression methods, retaining 94.7% of the original performance with only 128 tokens, achieving a 1.92times faster inference speed.","upvotes":6,"discussionId":"6a7181abec5082b9f872cf60","projectPage":"https://paper.pnu-cvsp.com/3DZip/","githubRepo":"https://github.com/cvsp-lab/3DZip","githubRepoAddedBy":"user","githubStars":13},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"66c2f277a2213740f48ff641","avatarUrl":"/avatars/118cd48dc1692db21f281b9a9f13d539.svg","isPro":false,"fullname":"changwoo baek","user":"higokri","type":"user"},{"_id":"696a2e2cb4cb5e359740af1d","avatarUrl":"/avatars/d40e6882a95ab92e99cfe29b98e07baf.svg","isPro":false,"fullname":"Beomkwon Cho","user":"chobk5194","type":"user"},{"_id":"651e1ac1bc5572510d79f29d","avatarUrl":"/avatars/2ed835dfe7732eb4ced77731fb1c9f58.svg","isPro":false,"fullname":"Sunchan Park","user":"sunchan","type":"user"},{"_id":"68395089a84097bd88377818","avatarUrl":"/avatars/caea9f7c70b1de757530e9127a15c93a.svg","isPro":false,"fullname":"Junsu Kim","user":"wave2503","type":"user"},{"_id":"6997f93bd14845772dfb4375","avatarUrl":"/avatars/7a6da7273f3689f7352eb79e0a0ec115.svg","isPro":false,"fullname":"KIMWOOHYEONG","user":"dngud1003","type":"user"},{"_id":"6a7185ab713a78042695e88c","avatarUrl":"/avatars/7e06cc85efc56c4a4344e06f5b4263dd.svg","isPro":false,"fullname":"Wooseok Seo (서우석)","user":"seo-wooseok","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.01185.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2608.01185 in a model README.md to link it from this page.

Cite arxiv.org/abs/2608.01185 in a dataset README.md to link it from this page.

Cite arxiv.org/abs/2608.01185 in a Space README.md to link it from this page.

Add this paper to a collection to link it from this page.