Latent distribution: Measuring floor plan typicality with isovist representation learning

Abstract

The effectiveness of machine learning (ML) models for architectural applications relies on high-quality datasets balanced with advancements in model architecture and computational capacity. Current methods for evaluating architectural floor plan datasets typically depend on explicit semantic annotations, which limit their effectiveness and scalability when annotations are unavailable or inconsistent. To address this limitation, this research develops an isovist-based latent representation approach to quantitatively measure typicality and diversity within architectural datasets without relying on semantic labels. We introduce Isovist Latent Norm Typicality, a metric that leverages the statistical structure of latent representations derived from isovist morphological features using a variational autoencoder (VAE). This metric quantifies typicality by analyzing distributional shifts in latent representations between individual floor plans and a reference dataset using a modified Wasserstein distance. Experimental results demonstrate the approach's ability to distinguish typical from atypical floor plan configurations, capturing the morphological features that complement conventional metrics. Comparative analysis indicates that our method provides insights into spatial organization, correlating with conventional properties such as programmatic diversity and spatial openness. By quantifying typicality through purely morphological features, the proposed methodology facilitates dataset curation prior to costly semantic annotation, enhancing dataset quality and enabling scalability to more extensive and diverse architectural datasets.

Keywords

Floor plan / Typicality / Isovist / Representation learning / Latent space / Dataset

References

[1] An, J. , Cho, S. , 2015. Variational autoencoder based anomaly detection using reconstruction probability. Spec. Lect. IE 2, 1- 18.

[2] Angiulli, F. , Fassetti, F. , Ferragina, L. , 2023. LatentOut: an unsupervised deep anomaly detection approach exploiting latent space distribution. Mach. Learn. 112, 4323- 4349.

[3] Asperti, A. , Trentin, M. , 2020. Balancing reconstruction error and kullback-leibler divergence in variational autoencoders. IEEE Access 8, 199440- 199448.

[4] Azizi, V. , Usman, M. , Sohn, S.S. , Schwartz, M. , Moon, S. , Faloutsos, P. , Kapadia, M. , 2023. The role of latent representations for design space exploration of floorplans. Simulation 99, 1167- 1179.

[5] Benedikt, M. , 1979. To take hold of space: isovists and isovist fields. Environ. Plann. Plann. Des. 6, 47- 65.

[6] Bengio, Y. , Courville, A. , Vincent, P. , 2013. Representation learning: a review and new perspectives. IEEE Trans. Pattern Anal. Mach. Intell. 35, 1798- 1828.

[7] Bhatt, Nikita , Bhatt, Nirav , Prajapati, P. , Sorathiya, V. , Alshathri, S. , El-Shafai, W. , 2024. A data-centric approach to improve performance of deep learning models. Sci. Rep. 14, 22329.

[8] Carpo, M. , 2024. Every dataset is a canon. Archit. Des. 94, 14- 19.

[9] Chaillou, S. , 2022. Latent typologies: architecture in latent space. In: Carta, S. (Ed.), Machine Learning and the City. John Wiley & Sons, Ltd, pp. 189-192.

[10] Chen, J. , Stouffs, R. , 2023. Deciphering the noisy landscape: architectural conceptual design space interpretation using disentangled representation learning. Comput. Aided Civ. Infrastruct. Eng. 38, 601- 620.

[11] Chen, C. , Wang, R. , Vogel, C. , Pollefeys, M. , 2024. F3Loc: fusion and filtering for floorplan localization. 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Presented at the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 18029-18038.

[12] Christenson, M. , 2014. Isovist-Based Occlusion Maps Representing Critical Variations in Le Corbusier's Museum of Unlimited Extension. Environ. Plann. Plann. Des. 41, 39- 52.

[13] Conroy Dalton, R. , Dalton, N.S. , McElhinney, S. , Mavros, P. , 2022. Isovist in a grid: benefits and limitations. In: van Nes, A., de Koning, R.E. (Eds.), Proceedings of the 13th Space Syntax Symposium. Western Norway University of Applied Sciences, Department of Civil Engineering, Bergen, Norway, p. 551.

[14] Dawes, M. , Ostwald, M.J. , 2013. Using isovists to analyse prospectrefuge theory: an examination of the usefulness of potential spatio-visual measures. Int. J. Constr. Environ. 3, 25- 40.

[15] de las Heras, L.-P. , Terrades, O.R. , Robles, S. , Sánchez, G. , 2015. CVC-FP and SGT: a new database for structural floor plan analysis and its groundtruthing tool. Int. J. Doc. Anal. Recognit. IJDAR 18, 15- 30.

[16] del Campo, M. , Manninger, S. , 2022. Architecture design in the age of artificial intelligence: the latent ontology of architectural features. In: The Routledge Companion to Ecological Design Thinking. Routledge, New York, pp. 75-91.

[17] Dong, J. , van Ameijde, J. , 2023. The impact of spatial layout on orientation and wayfinding in public housing estates using isovist polygons and shape matching algorithms. In: Turrin, M., Andriotis, C., Rafiee, A. (Eds.), Computer-Aided Architectural Design. INTERCONNECTIONS: Co-Computing Beyond Boundaries, Communications in Computer and Information Science. Springer Nature, Switzerland, Cham, pp. 655-669.

[18] Feld, S. , Illium, S. , Sedlmeier, A. , Belzner, L. , 2018. Trajectory annotation using sequences of spatial perception. In: Proceedings of the 26th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, SIGSPATIAL '18. ACM, New York, NY, USA, pp. 329-338.

[19] Goodfellow, I. , Pouget-Abadie, J. , Mirza, M. , Xu, B. , Warde-Farley, D. , Ozair, S. , Courville, A. , Bengio, Y. , 2014. Generative adversarial nets. In: Advances in Neural Information Processing Systems, pp. 2672-2680.

[20] Huszár, F. , 2017. Gaussian distributions are soap bubbles inFER-ENCe. inference.vc/high-dimensional-gaussian- distributionsare-soap-bubble/. (Accessed 18 June 2024).

[21] Johanes, M. , 2024. Floorplan isovist dataset.

[22] Johanes, M. , Huang, J. , 2023. Generative isovist transformer: inverted GANs for isovist representation in architectural floorplan. In: Pak, B., Wurzer, G., Stouffs, R. (Eds.), Co-Creating the Future: Inclusion in and Through Design-Proceedings of the 40th Conference on Education and Research in Computer Aided Architectural Design in Europe (eCAADe 2022). Ghent, Belgium, pp. 621-629.

[23] Johanes, M. , Huang, J. , 2022b. Latent isovist: discovery machine learning for spatial sequence synthesis. In: Dokonal, W., Hirschberg, U., Wurzer, G. (Eds.), Digital Design Reconsidered-Proceedings of the 41st Conference on Education and Research in Computer Aided Architectural Design in Europe (eCAADe 2023). Graz, Austria, pp. 471-480.

[24] Johanes, M. , Huang, J. , 2022a. Deep learning spatial signature: of machine-human interpretable spatial properties using inverted GANs. In: Akbarzadeh, M., Aviv, D., Jamelle, H., StuartSmith, R. (Eds.), ACADIA 2022: Hybrids and Haecceities[Proceedings of the 42nd Annual Conference of the Association of Computer Aided Design in Architecture (ACADIA). University of Pennsylvania Stuart Weitzman School of Design, Philadelphia, Pennsylvania, pp. 672-683.

[25] Johanes, M. , Huang, J. , 2021. Deep learning isovist: unsupervised spatial encoding in architecture. In: Bogosian, B., Dörfler, K., Farahi, B., Garcia del Castillo y López, J., Grant, J., Noel, Parascho, S., Scott (Eds.), ACADIA 2021: Realignments: toward Critical Computation[Proceedings of the 41st Annual Conference of the Association of Computer Aided Design in Architecture]. Online + Global, pp. 134-141.

[26] Kalervo, A. , Ylioinas, J. , Häikiö M. , Karhu, A. , Kannala, J. , 2019. CubiCasa5K: a dataset and an improved multi-task model for floorplan image analysis. In: Image Analysis: 21St Scandinavian Conference, SCIA 2019. Springer-Verlag, Norrköping, Sweden, pp. 28-40.

[27] Kilcher, Y. , Lucchi, A. , Hofmann, T. , 2018. Semantic Interpolation in Implicit Models. Presented at the International Conference on Learning Representations.

[28] Kim, F.C. , Yang, H. , Johanes, M. , Huang, J. , 2024. Deep winning form: machine investigation of architectural quality. In: ACCELERATED DESIGN-Proceedings of the 29th CAADRIA Conference, pp. 273-282. Singapore.

[29] Kingma, D.P. , Welling, M. , 2014. Auto-encoding variational bayes.In: ICLR 2014. Banff, Canada.

[30] Kohonen, T. , 1990. The self-organizing map. Proc. IEEE 78, 1464- 1480.

[31] Leduc, T. , Chaillou, F. , Ouard, T. , 2011. Towards a "typification" of the pedestrian surrounding space: analysis of the isovist using digital processing method. In: Advancing Geoinformation Science for a Changing World, Lecture Notes in Geoinformation and Cartography. Springer, Berlin, Heidelberg, pp. 275-292.

[32] Lee, P.K. , Stenger, B. , 2021. Shape-based floor plan retrieval using parse tree matching. In: 2021 17th International Conference on Machine Vision and Applications (MVA). Presented at the 2021 17th International Conference on Machine Vision and Applications (MVA), pp. 1-5.

[33] Liu, Y. , Jun, E. , Li, Q. , Heer, J. , 2019. Latent space cartography: visual analysis of vector space embeddings. Comput. Graph. Forum 38, 67- 78.

[34] McInnes, L. , Healy, J. , Saul, N. , Großberger, L. , 2018. UMAP: uniform manifold approximation and projection. J. Open Source Softw. 3, 861.

[35] Meng, S. , 2022. Exploring in the latent space of design: a method of plausible building facades images generation, properties control and model explanation base on StyleGAN2. In: Yuan, P.F., Chai, H., Yan, C., Leach, N. (Eds.), Proceedings of the 2021 DigitalFUTURES. Springer, Singapore, pp. 55-68.

[36] Patil, A.G. , Li, M. , Fisher, M. , Savva, M. , Zhang, H. , 2021. LayoutGMN: neural graph matching for structural layout similarity. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 11048-11057. Nashville, TN, USA.

[37] Peng, W. , Zhang, F. , Nagakura, T. , 2017. Machines' perception of space: employing 3D isovist methods and a convolutional neural network in architectural space classification. In: ACADIA 2017: DISCIPLINES & DISRUPTION[Proceedings of the 37th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA), pp. 474-481. Cambridge, MA.

[38] Pizarro, P.N. , Hitschfeld, N. , Sipiran, I. , 2023. Large-scale multiunit floor plan dataset for architectural plan analysis and recognition. Autom. Constr. 156, 105132.

[39] Rodrigues, E. , Sousa-Rodrigues, D. , Teixeira de Sampayo, M. , Gaspar, A.R. , Gomes, Á. , Henggeler Antunes, C. , 2017. Clustering of architectural floor plans: a comparison of shape representations. Autom. Constr. 80, 48- 65.

[40] Rumelhart, D.E. McClelland, J.L. , 1987. Learning internal representations by error propagation. In: Parallel Distributed Processing: Explorations in the Microstructure of Cognition: Foundations. MIT Press, pp. 318-362.

[41] Sedlmeier, A. , Feld, S. , 2018. Discovering and learning recurring structures in building floor plans. In: Kiefer, P., Huang, H., Van de Weghe, N., Raubal, M. (Eds.), Progress in Location Based Services 2018, Lecture Notes in Geoinformation and Cartography. Springer International Publishing, Cham, pp. 151-170.

[42] Sharma, D. , Gupta, N. , Chattopadhyay, C. , Mehta, S. , 2017. DANIEL: a deep architecture for automatic analysis and retrieval of building floor plans. In: 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), pp. 420-425.

[43] Shih, C.-H. , Peng, C.-H. , 2022. Floor plan exploration framework based on similarity distances. In: Smart Tools and Applications in Graphics-Eurographics Italian Chapter Conference. The Eurographics Association, pp. 115-117.

[44] Shim, J. , Moon, J. , Kim, H. , Hwang, E. , 2024. FloorDiffusion: diffusion model-based conditional floorplan image generation method using parameter-efficient fine-tuning and image inpainting. J. Build. Eng. 95, 110320.

[45] Son, K. , Hyun, K.H. , 2021. A framework for multivariate data based floor plan retrieval and generation. In: Globa, A., van Ameijde, J., Fingrut, A., Kim, N., Lo, T.T.S. (Eds.), PROJECTIONS-Proceedings of the 26th CAADRIA Conference. Chinese University of Hong Kong, Hong Kong, pp. 281-290.

[46] Stamps, A.E. , 2005. Isovists, enclosure, and permeability theory. Environ. Plann. Plann. Des. 32, 735- 762.

[47] Standfest, M. , Franzen, M. , Schröder, Y. , Medina, L.G. , Hernandez, Y.V. , Buck, J.H. , Tan, Y.-L. , Niedzwiecka, M. , Colmegna, R. , 2022. Swiss dwellings: a large dataset of apartment models including aggregated geolocation-based simulation results covering viewshed, natural light, traffic noise, centrality and geometric analysis.

[48] Tang, H. , Shao, L. , Sebe, N. , Van Gool, L. , 2024. Graph transformer GANs with graph masked modeling for architectural layout generation. IEEE Trans. Pattern Anal. Mach. Intell. 46, 4298- 4313.

[49] Triantafyllou, G. , Verbree, E. , Rafiee, A. , 2024. Indoor localisation through isovist fingerprinting from point clouds and floor plans.J. Locat. Based Serv. 1-20.

[50] van der Maaten, L. , Hinton, G. , 2008. Visualizing data using t-SNE. J. Mach. Learn. Res. 9, 2579- 2605.

[51] van Engelenburg, C.C.J. , Khademi, S. , Van Gemert, J.C. , 2023. SSIG: a visually-guided graph edit distance for floor plan similarity. In: 2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). IEEE, Paris, France, pp. 1565-1574.

[52] van Engelenburg, C. , Mostafavi, F. , Kuhn, E. , Jeon, Y. , Franzen, M. , Standfest, M. , van Gemert, J. , Khademi, S. , 2024. MSD: A Benchmark Dataset for Floor Plan Generation of Building Complexesin: Computer Vision-ECCV 2024: 18th European Conference, Milan, Italy, September 29-October 4, 2024, Proceedings, Part LVIII. Springer-Verlag, Berlin, Heidelberg, pp. 60-75.

[53] Wang, L. , Liu, J. , Zeng, Y. , Cheng, G. , Hu, H. , Hu, J. , Huang, X. , 2023. Automated building layout generation using deep learning and graph algorithms. Autom. Constr. 154, 105036.

[54] Wegner, S.-A. , 2024. Lecture notes on high-dimensional data.

[55] White, T. , 2016. Sampling generative networks.

[56] Wold, S. , Esbensen, K. , Geladi, P. , 1987. Principal component analysis. In: Chemom. Intell. Lab. Syst., Proceedings of the Multivariate Statistical Workshop for Geologists and Geochemists, vol. 2, pp. 37-52.

[57] Wonka, P. , Wimmer, M. , Sillion, F. , Ribarsky, W. , 2003. Instant architecture. ACM Trans. Graph. 22, 669- 677.

[58] Wu, W. , Fu, X.-M. , Tang, R. , Wang, Y. , Qi, Y.-H. , Liu, L. , 2019. Data-driven interior plan generation for residential buildings. ACM Trans. Graph. 38, 1- 12.

[59] Yao, R. , Liu, C. , Zhang, L. , Peng, P. , 2019. Unsupervised anomaly detection using variational auto-encoder based feature extraction. In: 2019 IEEE International Conference on Prognostics and Health Management (ICPHM). Presented at the 2019. IEEE International Conference on Prognostics and Health Management (ICPHM), pp. 1-7.

[60] Zhang, H. , Wang, P. , Li, M. , Li, Z. , Wu, Y. , 2025. Unit region encoding: a unified and compact geometry-aware representation for floorplan applications. CoRR abs/2501.11097.

RIGHTS & PERMISSIONS

The Author(s). Publishing services by Elsevier B.V. on behalf of Higher Education Press and KeAi.