PointInfinity: Resolution-Invariant Point Diffusion Models

We present PointInfinity, an efficient family of point cloud diffusion models. Our core idea is to use a transformer-based architecture with a fixed-size, resolution-invariant latent representation. This enables efficient training with low-resolution point clouds, while allowing high-resolution poin...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Proceedings (IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Online) s. 10050 - 10060
Hlavní autori: Huang, Zixuan, Johnson, Justin, Debnath, Shoubhik, Rehg, James M., Wu, Chao-Yuan
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 16.06.2024
Predmet:
ISSN:1063-6919
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract We present PointInfinity, an efficient family of point cloud diffusion models. Our core idea is to use a transformer-based architecture with a fixed-size, resolution-invariant latent representation. This enables efficient training with low-resolution point clouds, while allowing high-resolution point clouds to be generated during inference. More importantly, we show that scaling the test-time resolution beyond the training resolution improves the fidelity of generated point clouds and surfaces. We analyze this phenomenon and draw a link to classifier-free guidance commonly used in diffusion models, demonstrating that both allow trading off fidelity and variability during inference. Experiments on CO3D show that PointInfinity can efficiently generate high-resolution point clouds (up to 131k points, 31× more than Point-E) with state-of-the-art quality.
AbstractList We present PointInfinity, an efficient family of point cloud diffusion models. Our core idea is to use a transformer-based architecture with a fixed-size, resolution-invariant latent representation. This enables efficient training with low-resolution point clouds, while allowing high-resolution point clouds to be generated during inference. More importantly, we show that scaling the test-time resolution beyond the training resolution improves the fidelity of generated point clouds and surfaces. We analyze this phenomenon and draw a link to classifier-free guidance commonly used in diffusion models, demonstrating that both allow trading off fidelity and variability during inference. Experiments on CO3D show that PointInfinity can efficiently generate high-resolution point clouds (up to 131k points, 31× more than Point-E) with state-of-the-art quality.
Author Huang, Zixuan
Rehg, James M.
Wu, Chao-Yuan
Debnath, Shoubhik
Johnson, Justin
Author_xml – sequence: 1
  givenname: Zixuan
  surname: Huang
  fullname: Huang, Zixuan
  organization: FAIR at Meta
– sequence: 2
  givenname: Justin
  surname: Johnson
  fullname: Johnson, Justin
  organization: FAIR at Meta
– sequence: 3
  givenname: Shoubhik
  surname: Debnath
  fullname: Debnath, Shoubhik
  organization: FAIR at Meta
– sequence: 4
  givenname: James M.
  surname: Rehg
  fullname: Rehg, James M.
  organization: University of Illinois at Urbana-Champaign
– sequence: 5
  givenname: Chao-Yuan
  surname: Wu
  fullname: Wu, Chao-Yuan
  organization: FAIR at Meta
BookMark eNotj99KwzAcRqMoOGffYBd9gdZf_ie7k-q0MHEM9XYkTQKRmkrTCXt7i3r1XZyPA-caXaQheYRWGGqMQd8277s9J5LSmgBhNYDm6gwVWmpFOVBOAcQ5WmAQtBIa6ytU5PwBAJRgLLRaoPVuiGlqU4gpTqd1ufd56I9THFLVpm8zRpOm8vdT3scQjnkm5fPgfJ9v0GUwffbF_y7R2-bhtXmqti-PbXO3rSKWYqocNyC9lIQxzJ1yrAvgLbWGc-GYMNJaoqwKncDKMM47o1gnLPfBkblB0iVa_Xmj9_7wNcZPM54Oc9Ms4Jj-AK7rSzQ
CODEN IEEPAD
ContentType Conference Proceeding
DBID 6IE
6IH
CBEJK
RIE
RIO
DOI 10.1109/CVPR52733.2024.00958
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan (POP) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP) 1998-present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE/IET Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Applied Sciences
EISBN 9798350353006
EISSN 1063-6919
EndPage 10060
ExternalDocumentID 10655651
Genre orig-research
GroupedDBID 6IE
6IH
6IL
6IN
AAWTH
ABLEC
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
IEGSK
IJVOP
OCL
RIE
RIL
RIO
ID FETCH-LOGICAL-i176t-d5a07e7724415d8d4cf0eb3ba556d46a7bb28b8fc618a455ca84c6b5efd250373
IEDL.DBID RIE
ISICitedReferencesCount 1
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001342442401037&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
IngestDate Wed Aug 27 02:00:59 EDT 2025
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i176t-d5a07e7724415d8d4cf0eb3ba556d46a7bb28b8fc618a455ca84c6b5efd250373
PageCount 11
ParticipantIDs ieee_primary_10655651
PublicationCentury 2000
PublicationDate 2024-June-16
PublicationDateYYYYMMDD 2024-06-16
PublicationDate_xml – month: 06
  year: 2024
  text: 2024-June-16
  day: 16
PublicationDecade 2020
PublicationTitle Proceedings (IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Online)
PublicationTitleAbbrev CVPR
PublicationYear 2024
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0003211698
Score 2.2798243
Snippet We present PointInfinity, an efficient family of point cloud diffusion models. Our core idea is to use a transformer-based architecture with a fixed-size,...
SourceID ieee
SourceType Publisher
StartPage 10050
SubjectTerms 3D Diffusion Model
3D Generation
3D reconstruction
3D Vision
Computer vision
Deep Learning
Noise reduction
Point cloud compression
Shape
Three-dimensional displays
Training
Transformer cores
Title PointInfinity: Resolution-Invariant Point Diffusion Models
URI https://ieeexplore.ieee.org/document/10655651
WOSCitedRecordID wos001342442401037&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LSwMxEA5aPHiqj4pvcvC6urt591otFqQsotJbyWMCC7KV7rbgvzfZrtWLB29LWEiYZJLvm-SbQejGK5r7sO0nDKhPKFM0UdRFIKfAZ-CA2jaJ65OYTuVspopOrN5qYQCgfXwGt_Gzvct3C7uKobLg4ZwFABLIzq4QYiPW2gZUSKAyXMlOHpel6m70VjzH_GIk0MA8JslWsbD7ryIq7Rky7v-z9wM0-FHj4WJ7zhyiHaiOUL-Dj7hzzvoYDYtFWTWTypfBTT-HOEbmN-sqmVTrwImDEXH7D74vvV_FOBmOtdDe6wF6HT-8jB6TrjRCUmaCN4ljOhUQkHGkQ046an0aaLHRYYSOci2MyaWR3vJMasqY1ZJabhh4FzAPEeQE9apFBacIZ9KnhOSapwqoMkYBc4Ry6cHozDByhgbRFvOPTfaL-bcZzv9ov0D70dzxOVXGL1GvWa7gCu3ZdVPWy-t2zr4Apv6YBg
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LSwMxEA6igp7qo-LbPXhd3Wwem_RaLS3WskiV3kqymcCCbKXdFvz3JulavXjwtoSFhEkm-b5JvhmEbq2kqXXbfsyA2pgySWNJjQdyEiwGA7QISVyH2WgkJhOZN2L1oIUBgPD4DO78Z7jLN7Ni6UNlzsM5cwDEkZ0dRmmK13KtTUiFODLDpWgEcjiR9923_MVnGCOOCKY-Tbb0pd1_lVEJp0iv9c_-D1D7R48X5ZuT5hBtQXWEWg2AjBr3XByjTj4rq3pQ2dI56mcn8rH59cqKB9XKsWJnxij8Ez2U1i59pCzy1dDeF2302nscd_txUxwhLnHG69gwlWTgsLEnREYYWtjEEWOt3AgN5SrTOhVa2IJjoShjhRK04JqBNQ71kIycoO1qVsEpirCwCSGp4okEKrWWwAyhXFjQCmtGzlDb22L6sc5_Mf02w_kf7Tdorz9-Hk6Hg9HTBdr3pvePqzC_RNv1fAlXaLdY1eVifh3m7wvANptN
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=Proceedings+%28IEEE+Computer+Society+Conference+on+Computer+Vision+and+Pattern+Recognition.+Online%29&rft.atitle=PointInfinity%3A+Resolution-Invariant+Point+Diffusion+Models&rft.au=Huang%2C+Zixuan&rft.au=Johnson%2C+Justin&rft.au=Debnath%2C+Shoubhik&rft.au=Rehg%2C+James+M.&rft.date=2024-06-16&rft.pub=IEEE&rft.eissn=1063-6919&rft.spage=10050&rft.epage=10060&rft_id=info:doi/10.1109%2FCVPR52733.2024.00958&rft.externalDocID=10655651