Eliminating Rasterization: Direct Vector Floor Plan Generation With DiffPlanner

The boundary-constrained floor plan generation problem aims to generate the topological and geometric properties of a set of rooms within a given boundary. Recently, learning-based methods have made significant progress in generating realistic floor plans. However, these methods involve a workflow o...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE transactions on visualization and computer graphics Ročník 31; číslo 10; s. 7906 - 7922
Hlavní autoři: Wang, Shidong, Pajarola, Renato
Médium: Journal Article
Jazyk:angličtina
Vydáno: United States IEEE 01.10.2025
Témata:
ISSN:1077-2626, 1941-0506, 1941-0506
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract The boundary-constrained floor plan generation problem aims to generate the topological and geometric properties of a set of rooms within a given boundary. Recently, learning-based methods have made significant progress in generating realistic floor plans. However, these methods involve a workflow of converting vector data into raster images, using image-based generative models, and then converting the results back into vector data. This process is complex and redundant, often resulting in information loss. Raster images, unlike vector data, cannot scale without losing detail and precision. To address these issues, we propose a novel deep learning framework called DiffPlanner for boundary-constrained floor plan generation, which operates entirely in vector space. Our framework is a Transformer-based conditional diffusion model that integrates an alignment mechanism in training, aligning the optimization trajectory of the model with the iterative design processes of designers. This enables our model to handle complex vector data, better fit the distribution of the predicted targets, accomplish the challenging task of floor plan layout design, and achieve user-controllable generation. We conduct quantitative comparisons, qualitative evaluations, ablation experiments, and perceptual studies to evaluate our method. Extensive experiments demonstrate that DiffPlanner surpasses existing state-of-the-art methods in generating floor plans and bubble diagrams in the creative stages, offering more controllability to users and producing higher-quality results that closely match the ground truths.
AbstractList The boundary-constrained floor plan generation problem aims to generate the topological and geometric properties of a set of rooms within a given boundary. Recently, learning-based methods have made significant progress in generating realistic floor plans. However, these methods involve a workflow of converting vector data into raster images, using image-based generative models, and then converting the results back into vector data. This process is complex and redundant, often resulting in information loss. Raster images, unlike vector data, cannot scale without losing detail and precision. To address these issues, we propose a novel deep learning framework called DiffPlanner for boundary-constrained floor plan generation, which operates entirely in vector space. Our framework is a Transformer-based conditional diffusion model that integrates an alignment mechanism in training, aligning the optimization trajectory of the model with the iterative design processes of designers. This enables our model to handle complex vector data, better fit the distribution of the predicted targets, accomplish the challenging task of floor plan layout design, and achieve user-controllable generation. We conduct quantitative comparisons, qualitative evaluations, ablation experiments, and perceptual studies to evaluate our method. Extensive experiments demonstrate that DiffPlanner surpasses existing state-of-the-art methods in generating floor plans and bubble diagrams in the creative stages, offering more controllability to users and producing higher-quality results that closely match the ground truths.
The boundary-constrained floor plan generation problem aims to generate the topological and geometric properties of a set of rooms within a given boundary. Recently, learning-based methods have made significant progress in generating realistic floor plans. However, these methods involve a workflow of converting vector data into raster images, using image-based generative models, and then converting the results back into vector data. This process is complex and redundant, often resulting in information loss. Raster images, unlike vector data, cannot scale without losing detail and precision. To address these issues, we propose a novel deep learning framework called DiffPlanner for boundary-constrained floor plan generation, which operates entirely in vector space. Our framework is a Transformer-based conditional diffusion model that integrates an alignment mechanism in training, aligning the optimization trajectory of the model with the iterative design processes of designers. This enables our model to handle complex vector data, better fit the distribution of the predicted targets, accomplish the challenging task of floor plan layout design, and achieve user-controllable generation. We conduct quantitative comparisons, qualitative evaluations, ablation experiments, and perceptual studies to evaluate our method. Extensive experiments demonstrate that DiffPlanner surpasses existing state-of-the-art methods in generating floor plans and bubble diagrams in the creative stages, offering more controllability to users and producing higher-quality results that closely match the ground truths.The boundary-constrained floor plan generation problem aims to generate the topological and geometric properties of a set of rooms within a given boundary. Recently, learning-based methods have made significant progress in generating realistic floor plans. However, these methods involve a workflow of converting vector data into raster images, using image-based generative models, and then converting the results back into vector data. This process is complex and redundant, often resulting in information loss. Raster images, unlike vector data, cannot scale without losing detail and precision. To address these issues, we propose a novel deep learning framework called DiffPlanner for boundary-constrained floor plan generation, which operates entirely in vector space. Our framework is a Transformer-based conditional diffusion model that integrates an alignment mechanism in training, aligning the optimization trajectory of the model with the iterative design processes of designers. This enables our model to handle complex vector data, better fit the distribution of the predicted targets, accomplish the challenging task of floor plan layout design, and achieve user-controllable generation. We conduct quantitative comparisons, qualitative evaluations, ablation experiments, and perceptual studies to evaluate our method. Extensive experiments demonstrate that DiffPlanner surpasses existing state-of-the-art methods in generating floor plans and bubble diagrams in the creative stages, offering more controllability to users and producing higher-quality results that closely match the ground truths.
Author Wang, Shidong
Pajarola, Renato
Author_xml – sequence: 1
  givenname: Shidong
  orcidid: 0000-0003-2850-8319
  surname: Wang
  fullname: Wang, Shidong
  email: shwang@ifi.uzh.ch
  organization: Department of Informatics, University of Zurich, Zurich, Switzerland
– sequence: 2
  givenname: Renato
  orcidid: 0000-0002-6724-526X
  surname: Pajarola
  fullname: Pajarola, Renato
  email: pajarola@ifi.uzh.ch
  organization: Department of Informatics, University of Zurich, Zurich, Switzerland
BackLink https://www.ncbi.nlm.nih.gov/pubmed/40208765$$D View this record in MEDLINE/PubMed
BookMark eNpNkF1PwyAUhomZcR_6A0yM6aU3nRxKKXhn5jZNlsyYOS8b2p4qpmsndBf666VuGm9eDpznEHiGpFc3NRJyDnQMQNX1aj2Zjxll8TiKYyUkOyIDUBxCGlPR8zVNkpAJJvpk6Nw7pcC5VCekzymjMhHxgCynldmYWremfg2etGvRmi-_a-qb4M5YzNtg7aOxwaxqfD5Wug7mWKP9gYIX0755sCy7hj89JcelrhyeHdYReZ5NV5P7cLGcP0xuF2HOEtGGseQl5pwlwEAVQmAZaQXIWMaA0ixThSw0i0SUCaRQFD5lUYBWKP3PNUQjcrW_d2ubjx26Nt0Yl2PlX4HNzqURSCmBU648enlAd9kGi3RrzUbbz_RXggdgD-S2cc5i-YcATTvRaSc67USnB9F-5mI_YxDxH68EFTyOvgGHo3hx
CODEN ITVGEA
Cites_doi 10.1109/CVPR46437.2021.01342
10.1080/10447318.2020.1741118
10.1007/978-981-15-6568-7_8
10.1111/cgf.15007
10.1080/15230406.2021.1991478
10.1007/978-3-030-58452-8_10
10.1016/j.autcon.2013.06.005
10.1109/CVPR52688.2022.00764
10.1002/nav.3800020109
10.1145/3386569.3392391
10.1111/cgf.13380
10.1037/h0031619
10.5040/9781501325649
10.1145/3355089.3356556
10.1109/TVCG.2021.3126478
10.1109/ICCV.2017.241
10.1111/cgf.14984
10.1016/S0926-5805(00)00099-6
10.1109/CVPR.2017.632
10.1080/03052150214016
10.48550/ARXIV.1706.03762
10.1109/CVPR52729.2023.00529
10.1145/1882261.1866203
10.1145/3528223.3530135
10.1145/2461912.2461977
ContentType Journal Article
DBID 97E
RIA
RIE
AAYXX
CITATION
NPM
7X8
DOI 10.1109/TVCG.2025.3559682
DatabaseName IEEE All-Society Periodicals Package (ASPP) 2005–Present
IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE Electronic Library (IEL)
CrossRef
PubMed
MEDLINE - Academic
DatabaseTitle CrossRef
PubMed
MEDLINE - Academic
DatabaseTitleList
PubMed
MEDLINE - Academic
Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: RIE
  name: IEEE/IET Electronic Library (IEL) (UW System Shared)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
– sequence: 3
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISSN 1941-0506
EndPage 7922
ExternalDocumentID 40208765
10_1109_TVCG_2025_3559682
10960645
Genre orig-research
Journal Article
GroupedDBID ---
-~X
.DC
0R~
29I
4.4
53G
5GY
5VS
6IK
97E
AAJGR
AARMG
AASAJ
AAWTH
ABAZT
ABQJQ
ABVLG
ACGFO
ACIWK
AENEX
AETIX
AGQYO
AGSQL
AHBIQ
AI.
AIBXA
AKJIK
AKQYR
ALLEH
ALMA_UNASSIGNED_HOLDINGS
ATWAV
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
EBS
EJD
F5P
HZ~
H~9
IEDLZ
IFIPE
IFJZH
IPLJI
JAVBF
LAI
M43
O9-
OCL
P2P
PQQKQ
RIA
RIE
RNI
RNS
RZB
TN5
VH1
AAYXX
CITATION
NPM
7X8
ID FETCH-LOGICAL-c276t-584fec4271219d66ef3a91e22b2100bb9d8da2363b6e01dd6e08dd1a9e8110a13
IEDL.DBID RIE
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001566984900033&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1077-2626
1941-0506
IngestDate Sat Nov 01 14:35:31 EDT 2025
Sat Sep 06 11:15:50 EDT 2025
Sat Nov 29 07:35:16 EST 2025
Wed Sep 10 07:40:58 EDT 2025
IsPeerReviewed true
IsScholarly true
Issue 10
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
https://doi.org/10.15223/policy-029
https://doi.org/10.15223/policy-037
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c276t-584fec4271219d66ef3a91e22b2100bb9d8da2363b6e01dd6e08dd1a9e8110a13
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ORCID 0000-0003-2850-8319
0000-0002-6724-526X
PMID 40208765
PQID 3188814049
PQPubID 23479
PageCount 17
ParticipantIDs crossref_primary_10_1109_TVCG_2025_3559682
proquest_miscellaneous_3188814049
ieee_primary_10960645
pubmed_primary_40208765
PublicationCentury 2000
PublicationDate 2025-10-01
PublicationDateYYYYMMDD 2025-10-01
PublicationDate_xml – month: 10
  year: 2025
  text: 2025-10-01
  day: 01
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle IEEE transactions on visualization and computer graphics
PublicationTitleAbbrev TVCG
PublicationTitleAlternate IEEE Trans Vis Comput Graph
PublicationYear 2025
Publisher IEEE
Publisher_xml – name: IEEE
References Congalton (ref5) 1997; 63
ref13
ref12
ref14
ref31
ref30
ref11
ref32
Heusel (ref9)
ref1
Kingma (ref15) 2014
ref17
ref16
ref19
Dhariwal (ref6)
Ho (ref10)
Loshchilov (ref18)
ref24
ref23
Austin (ref2)
ref26
ref25
ref20
ref22
ref21
ref28
ref27
ref29
ref8
ref7
ref4
ref3
References_xml – start-page: 6840
  volume-title: Proc. Adv. Neural Inf. Process. Syst.
  ident: ref10
  article-title: Denoising diffusion probabilistic models
– ident: ref22
  doi: 10.1109/CVPR46437.2021.01342
– ident: ref26
  doi: 10.1080/10447318.2020.1741118
– ident: ref4
  doi: 10.1007/978-981-15-6568-7_8
– ident: ref12
  doi: 10.1111/cgf.15007
– year: 2014
  ident: ref15
  article-title: Adam: A method for stochastic optimization
– ident: ref13
  doi: 10.1080/15230406.2021.1991478
– ident: ref21
  doi: 10.1007/978-3-030-58452-8_10
– ident: ref24
  doi: 10.1016/j.autcon.2013.06.005
– start-page: 1
  volume-title: Proc. Int. Conf. Learn. Representations
  ident: ref18
  article-title: Decoupled weight decay regularization
– start-page: 8780
  volume-title: Proc. Adv. Neural Inf. Process. Syst.
  ident: ref6
  article-title: Diffusion models beat GANs on image synthesis
– start-page: 6629
  volume-title: Proc. Adv. Neural Inf. Process. Syst.
  ident: ref9
  article-title: GANs trained by a two time-scale update rule converge to a local nash equilibrium
– ident: ref8
  doi: 10.1109/CVPR52688.2022.00764
– ident: ref16
  doi: 10.1002/nav.3800020109
– ident: ref11
  doi: 10.1145/3386569.3392391
– ident: ref31
  doi: 10.1111/cgf.13380
– start-page: 17981
  volume-title: Proc. Adv. Neural Inf. Process. Syst.
  ident: ref2
  article-title: Structured denoising diffusion models in discrete state-spaces
– ident: ref7
  doi: 10.1037/h0031619
– volume: 63
  start-page: 425
  issue: 4
  year: 1997
  ident: ref5
  article-title: Exploring and evaluating the consequences of vector-to-raster and raster-to-vector conversion
  publication-title: Photogrammetric Eng. Remote Sens.
– ident: ref23
  doi: 10.5040/9781501325649
– ident: ref32
  doi: 10.1145/3355089.3356556
– ident: ref30
  doi: 10.1109/TVCG.2021.3126478
– ident: ref17
  doi: 10.1109/ICCV.2017.241
– ident: ref28
  doi: 10.1111/cgf.14984
– ident: ref1
  doi: 10.1016/S0926-5805(00)00099-6
– ident: ref14
  doi: 10.1109/CVPR.2017.632
– ident: ref20
  doi: 10.1080/03052150214016
– ident: ref29
  doi: 10.48550/ARXIV.1706.03762
– ident: ref25
  doi: 10.1109/CVPR52729.2023.00529
– ident: ref19
  doi: 10.1145/1882261.1866203
– ident: ref27
  doi: 10.1145/3528223.3530135
– ident: ref3
  doi: 10.1145/2461912.2461977
SSID ssj0014489
Score 2.4606364
Snippet The boundary-constrained floor plan generation problem aims to generate the topological and geometric properties of a set of rooms within a given boundary....
SourceID proquest
pubmed
crossref
ieee
SourceType Aggregation Database
Index Database
Publisher
StartPage 7906
SubjectTerms bubble diagram
Data models
deep generative modeling
Diffusion models
Floor plan generation
Floors
Generative adversarial networks
Iterative methods
Layout
Predictive models
Training
Transformers
Vectors
Title Eliminating Rasterization: Direct Vector Floor Plan Generation With DiffPlanner
URI https://ieeexplore.ieee.org/document/10960645
https://www.ncbi.nlm.nih.gov/pubmed/40208765
https://www.proquest.com/docview/3188814049
Volume 31
WOSCitedRecordID wos001566984900033&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
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE/IET Electronic Library (IEL) (UW System Shared)
  customDbUrl:
  eissn: 1941-0506
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0014489
  issn: 1077-2626
  databaseCode: RIE
  dateStart: 19950101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07T8MwED5BxQADzwLlUQWJCSklcR622VDVwoAKQqV0i5zYFpWqBvXB7-dsp1UZOrBYGew4urucv_O9AG4FZ4HSlPhhSAsf8X_h53GBg0qoNo4brW3J_Bfa67HhkL9Vyeo2F0YpZYPPVMs8Wl--LIuFuSrDP9zg7TjZhm1KU5estXIZ4D7cBRhSnyBMr1yYuOa-P2g_oSlIkhaerjxlpoWNsZtQEyR_ziPbYGUz1rRnTvfgn197CPsVuPQenTQcwZaaHMPeWsnBE3jtjG0fLxPt7L0LUyahysR88Jz68wb2It_rjkscTVMjz9WmNpO8z9H8CydqbbsdqWkdPrqdfvvZr5oq-AWh6dxHwKFVERMaoq6Saap0JHioCMnR-AvynEsmBYnSKE9VEEqJI5MyFFwxpKUIo1OoTcqJOgePxxFFAyfMBeKAQqBZLkkqdRCxWCZUsgbcLUmbfbvaGZm1OQKeGZZkhiVZxZIG1A0J1yY66jXgZsmNDCXfuDPERJWLWYbaiJl6XTFvwJlj02r1krsXG956CbtmcxeVdwW1-XShrmGn-JmPZtMmiteQNa14_QJoy8lW
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3JTsMwEB1BQQIO7EtYg8QJKSVxFtvcEKIUUQpCZblFTmwLpKpFJeX7GdspKgcOXKwc7MSamYzfeDaAE8FZqDQlQRTRMkD8XwZFUuKgUqqN40ZrWzK_Q7td9vrKH-pkdZsLo5SywWeqaR6tL18Oy7G5KsM_3ODtJJ2FuTRJSOjStX6cBvgl7kIMaUAQqNdOTFx11nu-vEZjkKRNPF95xkwTG2M5oS5If51ItsXK32jTnjqtlX_udxWWa3jpXzh5WIMZNViHpamigxtwf9W3nbxMvLP_KEyhhDoX89x3CtB_tlf5fqs_xNG0NfJddWozyX95r95wota235EabcJT66p32Q7qtgpBSWhWBQg5tCoTQiPUVjLLlI4FjxQhBZp_YVFwyaQgcRYXmQojKXFkUkaCK4a0FFG8BY3BcKB2wOdJTNHEiQqBSKAUaJhLkkkdxiyRKZXMg9MJafMPVz0jt1ZHyHPDktywJK9Z4sGmIeHUREc9D44n3MhR9o1DQwzUcPyZoz5ipmJXwj3Ydmz6WT3h7u4fbz2ChXbvrpN3brq3e7BoNuJi9PahUY3G6gDmy6_q_XN0aIXsG0-Hy7U
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%3Ajournal&rft.genre=article&rft.atitle=Eliminating+Rasterization%3A+Direct+Vector+Floor+Plan+Generation+With+DiffPlanner&rft.jtitle=IEEE+transactions+on+visualization+and+computer+graphics&rft.au=Wang%2C+Shidong&rft.au=Pajarola%2C+Renato&rft.date=2025-10-01&rft.eissn=1941-0506&rft.volume=31&rft.issue=10&rft.spage=7906&rft_id=info:doi/10.1109%2FTVCG.2025.3559682&rft_id=info%3Apmid%2F40208765&rft.externalDocID=40208765
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1077-2626&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1077-2626&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1077-2626&client=summon