pith. sign in

arxiv: 1907.00631 · v1 · pith:DW6SJJS6new · submitted 2019-07-01 · 💻 cs.GR

Automatic reconstruction of fully volumetric 3D building models from point clouds

Sebastian Ochmann , Richard Vock , Reinhard Klein This is my paper

Pith reviewed 2026-05-25 11:43 UTC · model grok-4.3

classification 💻 cs.GR
keywords 3D reconstructionpoint cloudsbuilding modelsinteger linear programmingvolumetric modelingindoor modelingroom segmentationparametric modeling
0
0 comments X

The pith

Integer linear programming reconstructs consistent volumetric 3D building models from unstructured indoor point clouds.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a method that reconstructs parametric, volumetric multi-story building models from raw unstructured point clouds. It first performs automatic room segmentation and outlier removal without relying on separate scans or prior segmentation. The core step casts reconstruction as an integer linear program over arrangements of volumetric wall entities, producing interconnected walls that fit the data exactly while enforcing hard constraints. A sympathetic reader would care because this removes common preprocessing requirements and yields usable 3D models for architecture or navigation from typical scan data.

Core claim

We present a novel method for reconstructing parametric, volumetric, multi-story building models from unstructured, unfiltered indoor point clouds by means of solving an integer linear optimization problem. Our approach overcomes limitations of previous methods by dropping assumptions about the input data such as the availability of separate scans as an initial room segmentation, instead performing fully automatic room segmentation and outlier removal. Restricting the solution space to arrangements of volumetric wall entities enforces a consistent model of volumetric, interconnected walls fitted to the observed data instead of unconnected, paper-thin surfaces. The optimization is formulated,

What carries the argument

Arrangements of volumetric wall entities fitted via integer linear programming

If this is right

  • Automatic room segmentation and outlier removal directly on unstructured point clouds without separate scans
  • Consistent models of interconnected volumetric walls instead of thin disconnected surfaces
  • Exact solutions to the optimization problem rather than approximations
  • Incorporation of hard constraints that were previously difficult to enforce
  • Demonstrated performance on a variety of complex real-world point clouds

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The parametric output models could be directly imported into building information modeling software for further editing.
  • If the integer linear program scales linearly with building size, the method could support city-scale indoor mapping projects.
  • Adding color or intensity attributes from the point cloud might allow the same framework to assign material properties to the walls.

Load-bearing premise

Building structure can be represented as arrangements of volumetric wall entities whose fitting via integer linear programming remains tractable on complex inputs.

What would settle it

A point cloud from a multi-story building where the integer linear program either returns no feasible solution or produces walls that leave large clusters of input points unexplained.

Figures

Figures reproduced from arXiv: 1907.00631 by Reinhard Klein, Richard Vock, Sebastian Ochmann.

Figure 1
Figure 1. Figure 1: Overview of the main steps. Ceiling of upper floor is hidden in (a), (c), and (f) for visualization purposes. (a) The input is a registered but otherwise [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Detected planes are the basis for our reconstruction. Di [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Dilation of surface support. Top: Example point cloud viewed from [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Explanation of notation and constraints. (a) Neighboring cells are considered as ordered pairs [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Our reconstruction result on the synthetic dataset “synth3” by the [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Comparison of a hand-crafted BIM model (left) and our recon [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Comparison between different reconstruction approaches. Left: Input point cloud viewed from above. Center: The method described in [5] may fail to regularize chains of almost coplanar walls, leading to segmented walls (circles). Also, reliance on separate scans as initial room labeling may lead to oversegmented rooms (dashed rectangle). Right: Our approach overcomes these issues by incorporating costs for … view at source ↗
Figure 8
Figure 8. Figure 8: Additional constraints may be added to interactively steer the re [PITH_FULL_IMAGE:figures/full_fig_p011_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: Different settings for the wall surface cost parameter α in Equation 7 demonstrated on the dataset “Case study 2” from the ISPRS Benchmark on Indoor Modeling. Center: Our default setting of α = 0.04 results in some walls to be fitted to windows which have high point support in this dataset. Also, a slab has a hole since floor support in staircases is often complex. Right: Increasing to α = 0.08 leads to s… view at source ↗
read the original abstract

We present a novel method for reconstructing parametric, volumetric, multi-story building models from unstructured, unfiltered indoor point clouds by means of solving an integer linear optimization problem. Our approach overcomes limitations of previous methods in several ways: First, we drop assumptions about the input data such as the availability of separate scans as an initial room segmentation. Instead, a fully automatic room segmentation and outlier removal is performed on the unstructured point clouds. Second, restricting the solution space of our optimization approach to arrangements of volumetric wall entities representing the structure of a building enforces a consistent model of volumetric, interconnected walls fitted to the observed data instead of unconnected, paper-thin surfaces. Third, we formulate the optimization as an integer linear programming problem which allows for an exact solution instead of the approximations achieved with most previous techniques. Lastly, our optimization approach is designed to incorporate hard constraints which were difficult or even impossible to integrate before. We evaluate and demonstrate the capabilities of our proposed approach on a variety of complex real-world point clouds.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper presents a method to reconstruct parametric, volumetric, multi-story 3D building models from unstructured indoor point clouds. It performs fully automatic room segmentation and outlier removal without requiring separate scans, represents building structure via arrangements of volumetric wall entities, formulates the fitting task as an integer linear program to obtain exact global solutions rather than approximations, incorporates hard constraints, and demonstrates results on complex real-world inputs.

Significance. If the ILP formulation remains tractable and produces exact optima on large unstructured inputs, the work would represent a meaningful advance by removing common input assumptions, enforcing volumetric consistency instead of thin surfaces, and enabling previously difficult hard constraints within an exact optimization framework.

major comments (1)
  1. [Abstract] Abstract: the central claim that the ILP 'allows for an exact solution instead of the approximations achieved with most previous techniques' while remaining practical for complex real-world point clouds is load-bearing, yet the abstract provides no bound on the number of binary variables (one per candidate wall entity) or on the growth of consistency constraints; because ILP is NP-hard, evidence is required that the variable-generation scheme prevents combinatorial explosion on noisy multi-story inputs.
minor comments (1)
  1. The abstract states that the approach is evaluated on a variety of complex real-world point clouds but does not name the datasets, report quantitative metrics (e.g., precision, recall, runtime), or describe how ground-truth models were obtained.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive review and positive assessment of the work's potential significance. We address the single major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that the ILP 'allows for an exact solution instead of the approximations achieved with most previous techniques' while remaining practical for complex real-world point clouds is load-bearing, yet the abstract provides no bound on the number of binary variables (one per candidate wall entity) or on the growth of consistency constraints; because ILP is NP-hard, evidence is required that the variable-generation scheme prevents combinatorial explosion on noisy multi-story inputs.

    Authors: We agree that the abstract does not provide explicit bounds on the number of binary variables or the growth of constraints. The candidate wall entities are generated from local geometric features extracted from the point cloud (detailed in Section 4), producing a number of variables that scales with the number of detected planar patches rather than all possible wall arrangements; consistency constraints are likewise generated only between spatially adjacent candidates. The manuscript contains no theoretical polynomial bound, as growth is data-dependent. However, Section 6 reports that all tested multi-story inputs (including noisy real-world scans with >50k points) produce ILP instances with at most a few thousand binary variables that solve to proven optimality in under 30 seconds using a standard solver. We will revise the abstract to include a brief reference to this observed practical scaling. revision: yes

Circularity Check

0 steps flagged

No circularity; derivation is self-contained optimization

full rationale

The paper defines an ILP over candidate volumetric wall entities, performs automatic segmentation on raw point clouds, and solves for a consistent volumetric model. No quoted step reduces a claimed prediction or uniqueness result to a fitted parameter or self-citation by construction. The central formulation takes external unstructured point clouds as input and applies standard ILP techniques; the derivation chain therefore remains independent of its own outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no details on free parameters, axioms, or invented entities; none can be identified from the given text.

pith-pipeline@v0.9.0 · 5702 in / 1050 out tokens · 27137 ms · 2026-05-25T11:43:11.740051+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

48 extracted references · 48 canonical work pages · 1 internal anchor

  1. [1]

    Planar 3D modeling of building interiors from point cloud data

    Sanchez, V , Zakhor, A. Planar 3D modeling of building interiors from point cloud data. In: Image Processing (ICIP), 2012 19th IEEE Interna- tional Conference on. IEEE; 2012, p. 1777–1780

    work page 2012

  2. [2]

    Indoor scene reconstruction using feature sensitive primitive extraction and graph-cut

    Oesau, S, Lafarge, F, Alliez, P. Indoor scene reconstruction using feature sensitive primitive extraction and graph-cut. ISPRS Journal of Photogrammetry and Remote Sensing 2014;90:68–82

    work page 2014

  3. [3]

    Fast, automated, scalable generation of textured 3D models of indoor environments

    Turner, E, Cheng, P, Zakhor, A. Fast, automated, scalable generation of textured 3D models of indoor environments. IEEE Journal of Selected Topics in Signal Processing 2015;9(3):409–421

    work page 2015

  4. [4]

    Piecewise-planar reconstruction of multi-room interiors with arbitrary wall arrangements

    Mura, C, Mattausch, O, Pajarola, R. Piecewise-planar reconstruction of multi-room interiors with arbitrary wall arrangements. In: Computer Graphics Forum; vol. 35. Wiley Online Library; 2016, p. 179–188

    work page 2016

  5. [5]

    Masset, R

    Ochmann, S, V ock, R, Wessel, R, Klein, R. Automatic reconstruction of parametric building models from indoor point clouds. Computers & Graphics 2016;54:94–103. URL: http://www.sciencedirect.com/ science/article/pii/S0097849315001119. doi:doi:10.1016/j. cag.2015.07.008; special Issue on CAD/Graphics 2015. Preprint / ISPRS Journal of Photogrammetry and Rem...

    work page doi:10.1016/j 2016

  6. [6]

    Indoor scan2bim: Building information models of house interiors

    Murali, S, Speciale, P, Oswald, MR, Pollefeys, M. Indoor scan2bim: Building information models of house interiors. In: Intelligent Robots and Systems (IROS), 2017 IEEE/RSJ International Conference on. IEEE; 2017, p. 6126–6133

    work page 2017

  7. [7]

    Floornet: A unified framework for floorplan reconstruction from 3D scans

    Liu, C, Wu, J, Furukawa, Y . Floornet: A unified framework for floorplan reconstruction from 3D scans. arXiv preprint arXiv:180400090 2018

    work page 2018

  8. [8]

    Automated 3D reconstruction of interiors from point clouds

    Budroni, A, Boehm, J. Automated 3D reconstruction of interiors from point clouds. International Journal of Architectural Computing 2010;8(1):55–73

    work page 2010

  9. [9]

    3D reconstruction of interior wall surfaces under oc- clusion and clutter

    Adan, A, Huber, D. 3D reconstruction of interior wall surfaces under oc- clusion and clutter. In: 3D Imaging, Modeling, Processing, Visualization and Transmission (3DIMPVT), 2011 International Conference on. IEEE; 2011, p. 275–281

    work page 2011

  10. [10]

    Automatic creation of se- mantically rich 3D building models from laser scanner data

    Xiong, X, Adan, A, Akinci, B, Huber, D. Automatic creation of se- mantically rich 3D building models from laser scanner data. Automation in Construction 2013;31:325–337

    work page 2013

  11. [11]

    Auto- matic room detection and reconstruction in cluttered indoor environments with complex room layouts

    Mura, C, Mattausch, O, Villanueva, AJ, Gobbetti, E, Pajarola, R. Auto- matic room detection and reconstruction in cluttered indoor environments with complex room layouts. Computers & Graphics 2014;44:20–32

    work page 2014

  12. [12]

    Towards automatic indoor reconstruction of cluttered building rooms from point clouds

    Previtali, M, Barazzetti, L, Brumana, R, Scaioni, M. Towards automatic indoor reconstruction of cluttered building rooms from point clouds. IS- PRS Annals of the Photogrammetry, Remote Sensing and Spatial Infor- mation Sciences 2014;2(5):281

    work page 2014

  13. [13]

    Automatic room segmentation from unstructured 3-d data of indoor environments

    Ambrus ¸, R, Claici, S, Wendt, A. Automatic room segmentation from unstructured 3-d data of indoor environments. IEEE Robotics and Au- tomation Letters 2017;2(2):749–756

    work page 2017

  14. [14]

    From point clouds to build- ing information models: 3D semi-automatic reconstruction of indoors of existing buildings

    Macher, H, Landes, T, Grussenmeyer, P. From point clouds to build- ing information models: 3D semi-automatic reconstruction of indoors of existing buildings. Applied Sciences 2017;7(10):1030

    work page 2017

  15. [15]

    Modeling indoor spaces using decomposition and reconstruction of structural elements

    Wang, R, Xie, L, Chen, D. Modeling indoor spaces using decomposition and reconstruction of structural elements. Photogrammetric Engineering & Remote Sensing 2017;83(12):827–841

    work page 2017

  16. [16]

    Toward automated modeling of floor plans

    Okorn, B, Xiong, X, Akinci, B, Huber, D. Toward automated modeling of floor plans. In: Proceedings of the symposium on 3D data processing, visualization and transmission; vol. 2. 2010,

    work page 2010

  17. [17]

    Raster-to-vector: Revisiting floorplan transformation

    Liu, C, Wu, J, Kohli, P, Furukawa, Y . Raster-to-vector: Revisiting floorplan transformation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017, p. 2195–2203

    work page 2017

  18. [18]

    Indoor scene reconstruction using primitive-driven space partitioning and graph-cut

    Oesau, S, Lafarge, F, Alliez, P. Indoor scene reconstruction using primitive-driven space partitioning and graph-cut. In: Eurographics work- shop on urban data modelling and visualisation. 2013,

    work page 2013

  19. [19]

    Reconstructing the worlds museums

    Xiao, J, Furukawa, Y . Reconstructing the worlds museums. International journal of computer vision 2014;110(3):243–258

    work page 2014

  20. [20]

    Robust reconstruction of interior building structures with multiple rooms under clutter and occlusions

    Mura, C, Mattausch, O, Villanueva, AJ, Gobbetti, E, Pajarola, R. Robust reconstruction of interior building structures with multiple rooms under clutter and occlusions. In: Computer-Aided Design and Computer Graphics (CAD/Graphics), 2013 International Conference on. IEEE; 2013, p. 52–59

    work page 2013

  21. [21]

    Reconstructing complex indoor environments with arbitrary wall orienta- tions

    Mura, C, Villanueva, AJ, Mattausch, O, Gobbetti, E, Pajarola, R. Reconstructing complex indoor environments with arbitrary wall orienta- tions. In: Eurographics (Posters). 2014, p. 19–20

    work page 2014

  22. [22]

    Floor plan generation and room labeling of indoor environments from laser range data

    Turner, E, Zakhor, A. Floor plan generation and room labeling of indoor environments from laser range data. In: Computer Graphics Theory and Applications (GRAPP), 2014 International Conference on. IEEE; 2014, p. 1–12

    work page 2014

  23. [23]

    Exploiting the room structure of buildings for scalable architectural modeling of interiors

    Mura, C, Pajarola, R. Exploiting the room structure of buildings for scalable architectural modeling of interiors. In: ACM SIGGRAPH 2017 Posters. ACM; 2017, p. 4

    work page 2017

  24. [24]

    Building modeling through enclosure reasoning

    Stambler, A, Huber, D. Building modeling through enclosure reasoning. In: 3D Vision (3DV), 2014 2nd International Conference on; vol. 2. IEEE; 2014, p. 118–125

    work page 2014

  25. [25]

    Automatic geometry generation from point clouds for BIM

    Thomson, C, Boehm, J. Automatic geometry generation from point clouds for BIM. Remote Sensing 2015;7(9):11753–11775

    work page 2015

  26. [26]

    Rapter: rebuilding man-made scenes with regular arrangements of planes

    Monszpart, A, Mellado, N, Brostow, GJ, Mitra, NJ. Rapter: rebuilding man-made scenes with regular arrangements of planes. ACM Trans Graph 2015;34(4):103–1

    work page 2015

  27. [27]

    Polyfit: Polygonal surface reconstruction from point clouds

    Nan, L, Wonka, P. Polyfit: Polygonal surface reconstruction from point clouds. In: The IEEE International Conference on Computer Vi- sion (ICCV). 2017,

    work page 2017

  28. [28]

    Automatic room segmentation of 3D laser data using morphological processing

    Jung, J, Stachniss, C, Kim, C. Automatic room segmentation of 3D laser data using morphological processing. ISPRS International Journal of Geo-Information 2017;6(7):206

    work page 2017

  29. [29]

    Room segmenta- tion in 3D point clouds using anisotropic potential fields

    Bobkov, D, Kiechle, M, Hilsenbeck, S, Steinbach, E. Room segmenta- tion in 3D point clouds using anisotropic potential fields. In: Multimedia and Expo (ICME), 2017 IEEE International Conference on. IEEE; 2017, p. 727–732

    work page 2017

  30. [30]

    Automatic generation of structural building descriptions from 3D point cloud scans

    Ochmann, S, V ock, R, Wessel, R, Tamke, M, Klein, R. Automatic generation of structural building descriptions from 3D point cloud scans. In: Computer Graphics Theory and Applications (GRAPP), 2014 Inter- national Conference on. IEEE; 2014, p. 1–8

    work page 2014

  31. [31]

    Classification of sensor independent point cloud data of building objects using random forests

    Bassier, M, Van Genechten, B, Vergauwen, M. Classification of sensor independent point cloud data of building objects using random forests. Journal of Building Engineering 2018

    work page 2018

  32. [32]

    Scan-to-bim for secondarybuilding components

    Ad ´an, A, Quintana, B, Prieto, SA, Bosch ´e, F. Scan-to-bim for secondarybuilding components. Advanced Engineering Informatics 2018;37:119–138

    work page 2018

  33. [33]

    Semantic as-built 3D modeling of structural elements of buildings based on local concavity and convexity

    Son, H, Kim, C. Semantic as-built 3D modeling of structural elements of buildings based on local concavity and convexity. Advanced Engineering Informatics 2017;34:114–124

    work page 2017

  34. [34]

    A framework for producing gbxml building geometry from point clouds for accurate and efficient building energy modelling

    Garwood, TL, Hughes, BR, O’Connor, D, Calautit, JK, Oates, MR, Hodgson, T. A framework for producing gbxml building geometry from point clouds for accurate and efficient building energy modelling. Applied Energy 2018;224:527–537

    work page 2018

  35. [35]

    Automatic open standard reporting for dimensional control compliance 2017

    Hyland, N, OKee ffe, S, Dore, C, Brodie, S. Automatic open standard reporting for dimensional control compliance 2017

    work page 2017

  36. [36]

    Automatic validation of as-is and as-generated ifc bims for advanced scan-to-bim methods 2017

    OKee ffe, S, Hyland, N, Dore, C, Brodie, S. Automatic validation of as-is and as-generated ifc bims for advanced scan-to-bim methods 2017

    work page 2017

  37. [37]

    The bim & scan R⃝ plat- form: A cloud-based cyber-physical system for automated solutions util- ising real & virtual worlds

    Brodie, S, Hyland, N, Dore, C, OKee ffe, S. The bim & scan R⃝ plat- form: A cloud-based cyber-physical system for automated solutions util- ising real & virtual worlds. 2017

    work page 2017

  38. [38]

    Data completion in building information management: electrical lines from range scans and photographs

    Krispel, U, Evers, HL, Tamke, M, Ullrich, T. Data completion in building information management: electrical lines from range scans and photographs. Visualization in Engineering 2017;5(1):4

    work page 2017

  39. [39]

    An ifc schema extension and binary serialization format to efficiently integrate point cloud data into building models

    Krijnen, T, Beetz, J. An ifc schema extension and binary serialization format to efficiently integrate point cloud data into building models. Ad- vanced Engineering Informatics 2017;33:473–490

    work page 2017

  40. [40]

    E fficient ransac for point-cloud shape detection

    Schnabel, R, Wahl, R, Klein, R. E fficient ransac for point-cloud shape detection. Computer Graphics Forum 2007;26(2):214–226

    work page 2007

  41. [41]

    A cluster algorithm for graphs

    Dongen, S. A cluster algorithm for graphs. Tech. Rep.; Amsterdam, The Netherlands, The Netherlands; 2000

    work page 2000

  42. [42]

    Point set shape detection

    Oesau, S, Verdie, Y , Jamin, C, Alliez, P, Lafarge, F, Giraudot, S. Point set shape detection. In: CGAL User and Reference Manual; 4.12 ed. CGAL Editorial Board; 2018,URL: https://doc.cgal.org/4.12/ Manual/packages.html#PkgPointSetShapeDetection3Summary

    work page 2018

  43. [43]

    Fast approximate energy minimiza- tion via graph cuts

    Boykov, Y , Veksler, O, Zabih, R. Fast approximate energy minimiza- tion via graph cuts. IEEE Transactions on pattern analysis and machine intelligence 2001;23(11):1222–1239

    work page 2001

  44. [44]

    Gurobi optimizer reference manual

    Gurobi Optimization, I. Gurobi optimizer reference manual. 2016. URL: http://www.gurobi.com

    work page 2016

  45. [45]

    3D is here: Point cloud library (pcl)

    Rusu, RB, Cousins, S. 3D is here: Point cloud library (pcl). In: Robotics and automation (ICRA), 2011 IEEE International Conference on. IEEE; 2011, p. 1–4

    work page 2011

  46. [46]

    2D arrangements

    Wein, R, Berberich, E, Fogel, E, Halperin, D, Hemmer, M, Salz- man, O, et al. 2D arrangements. In: CGAL User and Reference Manual; 4.12 ed. CGAL Editorial Board; 2018,URL: https://doc.cgal.org/ 4.12/Manual/packages.html#PkgArrangement2Summary

    work page 2018

  47. [47]

    The isprs benchmark on indoor modelling

    Khoshelham, K, Vilari ˜no, LD, Peter, M, Kang, Z, Acharya, D. The isprs benchmark on indoor modelling. International Archives of the Pho- togrammetry, Remote Sensing & Spatial Information Sciences 2017;42

    work page 2017

  48. [48]

    Joint 2D-3D-Semantic Data for Indoor Scene Understanding

    Armeni, I, Sax, A, Zamir, AR, Savarese, S. Joint 2D- 3D-Semantic Data for Indoor Scene Understanding. ArXiv e-prints 2017;arXiv:1702.01105

    work page internal anchor Pith review Pith/arXiv arXiv 2017