Pith. sign in

REVIEW 2 major objections 176 references

From one indoor photo, generative models can complete a metric floor map and put uncertainty where the layout is actually ambiguous.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-14 00:00 UTC pith:MGSYIAX3

load-bearing objection Solid new real-metric benchmark for single-view indoor BEV floor completion; generative models win under clean conditioning, but the paper already shows that edge largely vanishes with real monocular front-end noise. the 2 major comments →

arxiv 2603.16016 v2 pith:MGSYIAX3 submitted 2026-03-16 cs.CV cs.AIcs.ROeess.IV

FlatLands: Generative Floormap Completion From a Single Egocentric View

Subhransu S. Bhattacharjee , Dylan Campbell , Rahul Shome This is my paper
classification cs.CV cs.AIcs.ROeess.IV
keywords scene completionembodied AIgenerative modelingBEV floor mapsindoor navigationuncertaintyflow matchingegocentric vision
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

A robot or AR system that sees only a narrow egocentric photo still needs a usable floor plan of the room ahead: where it can walk, where walls and furniture block the way. FlatLands turns that need into a clean benchmark. It assembles 270,575 real metric observations from 17,656 indoor scenes, each with a bird's-eye view of what the camera already sees, what remains hidden, and the true full floor. On this testbed the authors show that stochastic generative models—diffusion and especially flow matching—recover more accurate completions than single-shot networks or seed ensembles, and that their sample-to-sample disagreement concentrates at doorways and room boundaries rather than filling geometrically fixed interiors with noise. The same ranking holds on held-out buildings never seen in training, and a full monocular RGB pipeline confirms the idea still works when depth and floor segmentation must be estimated rather than given. The result is a standardized place to measure uncertainty-aware floor completion for indoor navigation.

Core claim

Conditional generative models trained to sample multiple plausible bird's-eye floor completions from a single partial observation outperform deterministic predictors and epistemic ensembles both in fidelity and in calibrated diversity; in particular, flow matching with cross-attention localizes per-pixel variance to structurally ambiguous boundaries while ensembles spread variance uniformly across interiors that should be certain.

What carries the argument

FlatLands: a large multi-source dataset of aligned binary BEV maps (observed floor, unobserved mask, ground-truth floor, validity) together with the Masked Energy Score that jointly scores how close samples are to the true completion and how diverse they are among themselves.

Load-bearing premise

The ranking of completers is measured mainly on clean bird's-eye maps taken from perfect 3D meshes; once real monocular depth and segmentation noise enter, the practical advantage of generative sampling largely disappears.

What would settle it

On a large set of real monocular RGB frames with the same front-end depth and floor segmentation used in the paper, measure whether the best generative model still achieves reliably lower Masked Energy Score and higher best-of-K IoU than the best deterministic U-Net; if the gap vanishes or reverses, the main claim does not transfer to the intended sensor setting.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Navigation planners can treat multi-sample floor completions as probabilistic spatial priors and prefer actions that reduce high-variance boundary cells.
  • Epistemic seed ensembles are a poor stand-in for aleatoric layout ambiguity; conditional generative models should be preferred when uncertainty maps must be trustworthy.
  • The same conditioning interface (observed floor plus unobserved mask) can be reused for multi-view fusion or multi-floor buildings without redesigning the completion stage.
  • Because the target is an abstract binary metric grid, any front-end that produces that pair can plug into the same completers, making the stage sensor-agnostic.

Where Pith is reading between the lines

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

  • Improving monocular depth and floor segmentation will currently buy more navigation performance than further refining the generative completer, because the end-to-end bottleneck has already shifted upstream.
  • The same boundary-concentrated variance pattern should appear in outdoor bird's-eye map completion if the generator is properly conditioned on partial occupancy rather than texture.
  • Closed-loop exploration policies that actively drive the camera toward high-variance boundary cells would provide a direct test of whether calibrated floor uncertainty improves real-world coverage efficiency.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 0 minor

Summary. The paper formalizes single-view indoor BEV floor completion as a conditional inverse problem: from one egocentric RGB image, recover a distribution over binary traversability maps on the unobserved region while exactly preserving observed evidence. It releases FlatLands (270,575 observations from 17,656 real metric scenes across six sources), with aligned Fobs, U, V, and F* maps, scene-disjoint ID/OOD splits (ScanNet++ held out), and a shared evaluation protocol. Eleven methods are compared under identical (Fobs, U) conditioning, hard evidence clamping, and Reval masking, spanning naive fills, deterministic U-Net/PConv/LaMa, a LaMa seed ensemble, and diffusion/flow-matching generators (including FM+XAttn). New metrics UMR and Masked Energy Score (MES) are introduced; a boundary-variance decomposition and an end-to-end monocular RGB pipeline are also reported. The central empirical claim is that stochastic generators, especially FM+XAttn, better capture observation-conditioned uncertainty than point estimates or seed ensembles, with lower MES and variance concentrated at layout boundaries.

Significance. If the ranking and calibration findings hold, FlatLands supplies a missing standardized real-world testbed for uncertainty-aware indoor BEV completion, which is directly relevant to navigation, belief-space planning, and active perception. Strengths include the large multi-source metric corpus with explicit ID/OOD partitioning, a carefully shared conditioning and clamping protocol, oracle-free MES alongside best-of-K, the boundary-interior variance analysis (Fig. 4, Sec. S3), difficulty-tier and hard-subset ablations, a multi-solution synthetic case, and public release of data, code, and weights. The work cleanly isolates completion quality from front-end estimation while still reporting end-to-end transfer, which is valuable for the community even if practical gains shrink under noisy conditioning.

major comments (2)
  1. Sec. 6 and Tab. S1 / Sec. S4.1: The main-table ranking (Tabs. 3–4) is obtained under mesh-derived GT BEV conditioning. The paper’s own end-to-end RGB results show that the stochastic advantage over the best deterministic model largely vanishes once DepthPro/SegFormer noise is present, and the bottleneck shifts to perception. This is disclosed but under-emphasized relative to the abstract and conclusion claims about generative completion for embodied navigation. The camera-ready version should state more prominently (abstract/conclusion and main results discussion) that the GT-conditioned ranking is a controlled diagnostic of completion quality, not a direct claim of end-to-end navigation gain, and should report the end-to-end numbers in the main body rather than only the supplement.
  2. Sec. 5.2 and Tab. 3: High floor prevalence on Reval (mean ≈0.80) makes All-Floor competitive with LaMa on IoU/F1 and can inflate absolute scores. The hard subset (Sec. S7.4, Tab. S11) and Learnable-tier breakdown (Tabs. S8–S9) correctly address this, but they remain supplementary. Elevating at least one prevalence-controlled or low-rcond table into the main results would make the ranking less sensitive to label imbalance and better support the claim that generative models help where the task is genuinely hard.

Circularity Check

0 steps flagged

No circularity: FlatLands is an empirical benchmark whose rankings and uncertainty claims are scored against held-out mesh ground truth under a fixed protocol, not quantities defined from the models' own losses or self-cited uniqueness theorems.

full rationale

The paper formalizes single-view BEV floor completion, releases FlatLands (270k observations from six external real-metric sources with scene-disjoint ID/OOD splits and ScanNet++ held out), and ranks training-free, deterministic, ensemble, and stochastic completers under a shared interface (Fobs, U), hard evidence clamping (Eq. 3), and evaluation only on Reval = U ⊙ V. Fidelity metrics (UMR, IoU, F1) and the Masked Energy Score (MES) are computed against independent mesh-derived ground-truth floormaps F⋆; MES is the standard energy score with Jaccard distance and is not a re-labeling of the training objective. Boundary-variance localization (Fig. 4, Sec. S3) is a post-hoc diagnostic on sample variance, not a fitted prediction. End-to-end RGB results (Tab. S1) use external DepthPro/SegFormer front-ends and still score against mesh GT. Self-citations (e.g. authors' related arXiv notes) are not load-bearing for the ranking or for any uniqueness claim. No step reduces a claimed prediction to a fitted constant or to a definitional identity. Score 0 is the correct outcome for this class of paper.

Axiom & Free-Parameter Ledger

6 free parameters · 5 axioms · 4 invented entities

The central empirical claim rests on a constructed observation model (mesh visibility + ray occlusion), a binary traversability abstraction, several hand-chosen filtering and evaluation thresholds, and the assumption that masked generative training with hard clamping approximates the desired posterior. No new physical entities are postulated; the invented pieces are the dataset, metrics, and a conditioning architecture variant.

free parameters (6)
  • conditioning-signal threshold τ (rcond ≥ 0.1)
    Single hyperparameter that drops ~36% of synthesized observations; chosen by retention-curve inspection (Sec. S2.4–S2.5). Directly shapes the learnable regime used for all rankings.
  • sample count K=4
    Fixed for all stochastic methods and MES; sensitivity shown only up to K=4 (Sec. S7.2). Affects diversity and best-of-K scores.
  • classifier-free guidance scale s=2.0
    Selected as joint optimum on OOD IoU (Tab. S4); under- or over-guidance changes ranking margins.
  • boundary radius r=7 px for variance decomposition
    Defines interior vs boundary sets used to claim FM+XAttn localizes uncertainty (Sec. S3). Qualitative conclusion stable for r∈{5…9} but the numeric ratio depends on r.
  • BEV grid resolution Δ≈0.039 m/px (256×256)
    Canonical release resolution after downsampling from 0.01 m/px synthesis; trades doorway fidelity against memory (Sec. S2.3). All metrics inherit this discretization.
  • binarization threshold 0.5
    Fixed map threshold applied to all continuous outputs before scoring (Sec. 5.1–5.2).
axioms (5)
  • domain assumption Binary floor/non-floor on a metric BEV grid is a sufficient traversability primitive for collision checking and path planning.
    Stated in Sec. 3; richer semantics and graded traversability are declared orthogonal. If false for many platforms, the benchmark's ranking may not transfer.
  • domain assumption Mesh-derived visibility (FOV + ray occlusion) plus hard evidence clamping correctly defines the support of the observation-conditioned posterior.
    Eqs. (2)–(3) and Sec. 3.3; main tables evaluate under this synthetic conditioning.
  • domain assumption Indoor layouts exhibit structural regularities that generative models can internalize for multi-modal completion.
    Motivation in Sec. 1–2; underpins the claim that stochastic generators are preferable to point estimates.
  • ad hoc to paper Masked BCE (or reconstruction+adversarial) on Reval = U ⊙ V plus CFG yields samples from a useful approximation to P(Funo | Fobs, U).
    Training and sampling protocol in Sec. 5.1 and S5; not derived from a formal posterior guarantee.
  • ad hoc to paper Jaccard distance on Reval is a proper base metric for the Masked Energy Score ranking of multi-hypothesis completers.
    Eq. (5) and Sec. S6; propriety argued via Steinhaus transformation, but the ranking still depends on floor prevalence and mask construction.
invented entities (4)
  • FlatLands dataset and ID/OOD benchmark protocol no independent evidence
    purpose: Provide aligned real-metric single-view BEV floor observations with visibility, validity, and held-out ScanNet++ OOD split.
    Core contribution; independent evidence will exist once the promised public release is available and used by others.
  • Unobserved-region mismatch rate (UMR) no independent evidence
    purpose: Prevalence-aware fidelity metric on Reval that is less inflated by high floor prior than IoU alone.
    Defined in Sec. 5.2; useful but not independently validated outside this paper.
  • Masked Energy Score (MES) with Jaccard base no independent evidence
    purpose: Joint fidelity–diversity score for multi-sample completers under the evaluation mask.
    Adaptation of energy score (Gneiting & Raftery) to masked BEV maps; ranking tool for the paper's stochastic comparison.
  • FM+XAttn (flow matching with coarse cross-attention conditioning) no independent evidence
    purpose: Architectural variant that injects a condition encoder via cross-attention at 64×64 and 32×32.
    Modular conditioning; ablation shows most gain comes from the encoder rather than attention itself (Tab. S5).

pith-pipeline@v1.1.0-grok45 · 40029 in / 4012 out tokens · 50625 ms · 2026-07-14T00:00:54.355681+00:00 · methodology

0 comments
read the original abstract

A single egocentric image typically captures only a small portion of the floor, yet a complete metric traversability map of the surroundings would better serve applications such as indoor navigation. We introduce FlatLands, a dataset and benchmark for single-view bird's-eye view (BEV) floor completion. The dataset contains 270,575 observations from 17,656 real metric indoor scenes drawn from six existing datasets, with aligned observation, visibility, validity, and ground-truth BEV maps, and the benchmark includes both in- and out-of-distribution evaluation protocols. We compare training-free approaches, deterministic models, ensembles, and stochastic generative models. Finally, we instantiate the task as an end-to-end monocular RGB-to-floormaps pipeline. FlatLands provides a rigorous testbed for uncertainty-aware indoor mapping and generative completion for embodied navigation.

discussion (0)

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

Reference graph

Works this paper leans on

176 extracted references · 10 linked inside Pith

  1. [1]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2021)

    Austin,J.,Johnson,D.D.,Ho,J.,Tarlow,D.,vandenBerg,R.:StructuredDenois- ing Diffusion Models in Discrete State-Spaces. In: Advances in Neural Information Processing Systems (NeurIPS) (2021)

    2021

  2. [2]

    Axelrod, B., Kaelbling, L.P., Lozano-Pérez, T.: Provably Safe Robot Navigation withObstacleUncertainty.TheInternationalJournalofRoboticsResearch(2018)

    2018

  3. [3]

    In: Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2012)

    Aydemir, A., Jensfelt, P., Folkesson, J.: What Can We Learn from 38,000 Rooms? Reasoning about Unexplored Space in Indoor Environments. In: Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2012)

    2012

  4. [4]

    In: Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2025)

    Baek, S., Moon, B., Kim, S., Cao, M., Ho, C., Scherer, S., Jeon, J.: PIPE Planner: Pathwise information gain with map predictions for indoor robot exploration. In: Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2025)

    2025

  5. [5]

    Banfi, J., Woo, L., Campbell, M.: Is it Worth to Reason about Uncertainty in Occupancy Grid Maps during Path Planning? In: Proceedings of the IEEE Inter- national Conference on Robotics and Automation (ICRA) (2022)

    2022

  6. [6]

    In: International Joint Conference on Artificial Intelligence (IJCAI) (1977)

    Barrow, H.G., Tenenbaum, J.M., Bolles, R.C., Wolf, H.C.: Parametric Corre- spondence and Chamfer Matching: Two New Techniques for Image Matching. In: International Joint Conference on Artificial Intelligence (IJCAI) (1977)

    1977

  7. [7]

    In: Advances in Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track (2021)

    Baruch, G., Chen, Z., Dehghan, A., Dimry, T., Feigin, Y., Fu, P., Gebauer, T., Kurz, D., Joffe, B., Schwartz, A., Shulman, E.: ARKitScenes: A diverse real- world dataset for 3D indoor scene understanding using mobile RGB-D data. In: Advances in Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track (2021)

    2021

  8. [8]

    In: Proceedings of the IEEE/CVF International Conference on Com- puter Vision (ICCV) (2019)

    Behley, J., Garbade, M., Milioto, A., Quenzel, J., Behnke, S., Stachniss, C., Gall, J.: SemanticKITTI: A dataset for semantic scene understanding of LiDAR se- quences. In: Proceedings of the IEEE/CVF International Conference on Com- puter Vision (ICCV) (2019)

    2019

  9. [9]

    arXiv preprint arXiv:1308.3432 (2013)

    Bengio, Y., Léonard, N., Courville, A.: Estimating or Propagating Gradi- ents Through Stochastic Neurons for Conditional Computation. arXiv preprint arXiv:1308.3432 (2013)

    Pith/arXiv arXiv 2013

  10. [10]

    In: Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI) (2019)

    Bercher, P., Alford, R., Höller, D.: A Survey on Hierarchical Planning – One Abstract Idea, Many Concrete Realizations. In: Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI) (2019)

    2019

  11. [11]

    In: Pro- ceedings of the 27th Annual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH) (2000)

    Bertalmio, M., Sapiro, G., Caselles, V., Ballester, C.: Image Inpainting. In: Pro- ceedings of the 27th Annual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH) (2000)

    2000

  12. [12]

    arXiv preprint arXiv:2510.11014 (2025)

    Bhattacharjee, S.S., Lu, H., Campbell, D., Shome, R.: Into the Unknown: To- wards using generative models for sampling priors of environment uncertainty for planning in configuration spaces. arXiv preprint arXiv:2510.11014 (2025)

    Pith/arXiv arXiv 2025

  13. [13]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2025)

    Bhattacharjee, S.S., Campbell, D., Shome, R.: Believing is Seeing: Unobserved Object Detection using Generative Models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2025)

    2025

  14. [14]

    Next-Best-View

    Bircher, A., Kamel, M., Alexis, K., Oleynikova, H., Siegwart, R.: Receding Hori- zon "Next-Best-View" Planner for 3D Exploration. In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2016)

    2016

  15. [15]

    C., Lorenz, D., Müller, J., Podell, D., Rombach, R., Saini, H., Sauer, A., Smith, L.: FLUX.1 Kontext: Flow matching for in-context image generation and editing in latent space

    Black Forest Labs, Batifol, S., Blattmann, A., Boesel, F., Consul, S., Diagne, C., Dockhorn, T., English, J., English, Z., Esser, P., Kulal, S., Lacey, K., Levi, Y., Li, 16 Bhattacharjee et al. C., Lorenz, D., Müller, J., Podell, D., Rombach, R., Saini, H., Sauer, A., Smith, L.: FLUX.1 Kontext: Flow matching for in-context image generation and editing in ...

    Pith/arXiv arXiv 2025

  16. [16]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2018)

    Blau, Y., Michaeli, T.: The Perception-Distortion Tradeoff. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2018)

    2018

  17. [17]

    In: International Conference on Learning Representations (ICLR) (2025)

    Bochkovskii, A., Delaunoy, A., Germain, H., Santos, M., Zhou, Y., Richter, S.R., Koltun, V.: Depth Pro: Sharp Monocular Metric Depth in Less Than a Second. In: International Conference on Learning Representations (ICLR) (2025)

    2025

  18. [18]

    Numerische Mathematik (1991)

    Bunse-Gerstner, A., Byers, R., Mehrmann, V., Nichols, N.K.: Numerical compu- tation of an analytic singular value decomposition of a matrix valued function. Numerische Mathematik (1991)

    1991

  19. [19]

    IEEE Transactions on Robotics (2016)

    Cadena, C., Carlone, L., Carrillo, H., Latif, Y., Scaramuzza, D., Neira, J., Reid, I., Leonard, J.J.: Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age. IEEE Transactions on Robotics (2016)

    2016

  20. [20]

    In: Pro- ceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2023)

    Carvalho, J., Le, A.T., Baierl, M., Koert, D., Peters, J.: Motion Planning Diffu- sion: Learning and Planning of Robot Motions with Diffusion Models. In: Pro- ceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2023)

    2023

  21. [21]

    IEEE Transactions on Robotics (2025)

    Carvalho, J., Le, A.T., Kicki, P., Koert, D., Peters, J.: Motion Planning Diffusion: Learning and Adapting Robot Motion Planning with Diffusion Models. IEEE Transactions on Robotics (2025)

    2025

  22. [22]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023)

    Chan, E.R., Nagano, K., Chan, M.A., Bergman, A.W., Park, J.J., Levy, A., Ait- tala, M., Mello, S.D., Karras, T., Wetzstein, G.: Generative Novel View Synthesis with 3D-Aware Diffusion Models. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023)

    2023

  23. [23]

    Journal of Visual Communication and Image Representation (2001)

    Chan, T.F., Shen, J.: Nontexture Inpainting by Curvature-Driven Diffusions. Journal of Visual Communication and Image Representation (2001)

    2001

  24. [24]

    In: International Conference on 3D Vision (3DV) (2017)

    Chang, A.X., Dai, A., Funkhouser, T., Halber, M., Nießner, M., Savva, M., Song, S., Zeng, A., Zhang, Y.: Matterport3D: Learning from RGB-D data in indoor environments. In: International Conference on 3D Vision (3DV) (2017)

    2017

  25. [25]

    In: Proceedings of the IEEE/CVF In- ternational Conference on Computer Vision (ICCV) (2019)

    Chen, J., Liu, C., Wu, J., Furukawa, Y.: Floor-SP: Inverse CAD for Floorplans by Sequential Room-wise Shortest Path. In: Proceedings of the IEEE/CVF In- ternational Conference on Computer Vision (ICCV) (2019)

    2019

  26. [26]

    arXiv preprint arXiv:1604.06174 (2016)

    Chen, T., Xu, B., Zhang, C., Guestrin, C.: Training Deep Nets with Sublinear Memory Cost. arXiv preprint arXiv:1604.06174 (2016)

    Pith/arXiv arXiv 2016

  27. [27]

    In: International Conference on Learning Representations (ICLR) (2023)

    Chen, T., Zhang, R., Hinton, G.: Analog bits: Generating discrete data using diffusion models with self-conditioning. In: International Conference on Learning Representations (ICLR) (2023)

    2023

  28. [28]

    In: Proceedings of the International Conference on Machine Learning (ICML) (2018)

    Chen, X., Mishra, N., Rohaninejad, M., Abbeel, P.: PixelSNAIL: An improved autoregressive generative model. In: Proceedings of the International Conference on Machine Learning (ICML) (2018)

    2018

  29. [29]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    Cheng, B., Girshick, R., Dollár, P., Berg, A.C., Kirillov, A.: Boundary IoU: Im- proving object-centric image segmentation evaluation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    2021

  30. [30]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022) Generative Single-View Floormap Completion 17

    Cheng, Y.C., Lin, C.H., Lee, H.Y., Ren, J., Tulyakov, S., Yang, M.H.: In&out: Diverse image outpainting via GAN inversion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022) Generative Single-View Floormap Completion 17

    2022

  31. [31]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2020)

    Chi, L., Jiang, B., Mu, Y.: Fast Fourier Convolution. In: Advances in Neural Information Processing Systems (NeurIPS) (2020)

    2020

  32. [32]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2024)

    Christopher, J.K., Baek, S., Fioretto, F.: Constrained Synthesis with Pro- jected Diffusion Models. In: Advances in Neural Information Processing Systems (NeurIPS) (2024)

    2024

  33. [33]

    In: International Conference on Learning Representations (ICLR) (2023)

    Chung, H., Kim, J., McCann, M.T., Klasky, M.L., Ye, J.C.: Diffusion Posterior Sampling for General Noisy Inverse Problems. In: International Conference on Learning Representations (ICLR) (2023)

    2023

  34. [34]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    Cruz, S., Hutchcroft, W., Li, Y., Khosravan, N., Boyadzhiev, I., Kang, S.B.: Zillow Indoor Dataset: Annotated floor plans with 360◦ panoramas and 3D room layouts. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    2021

  35. [35]

    In: Robotics: Science and Systems (RSS) (2024)

    Curtis, A., Matheos, G., Gothoskar, N., Mansinghka, V., Tenenbaum, J.B., Lozano-Pérez, T., Kaelbling, L.P.: Partially Observable Task and Motion Plan- ning with Uncertainty and Risk Awareness. In: Robotics: Science and Systems (RSS) (2024)

    2024

  36. [36]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017)

    Dai, A., Chang, A.X., Savva, M., Halber, M., Funkhouser, T., Nießner, M.: Scan- Net: Richly-annotated 3D reconstructions of indoor scenes. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017)

    2017

  37. [37]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2021)

    Dhariwal, P., Nichol, A.: Diffusion Models Beat GANs on Image Synthesis. In: Advances in Neural Information Processing Systems (NeurIPS) (2021)

    2021

  38. [38]

    Ecol- ogy (1945)

    Dice, L.R.: Measures of the amount of ecologic association between species. Ecol- ogy (1945)

    1945

  39. [39]

    In: International Conference on Learning Representations (ICLR) (2021)

    Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., Houlsby, N.: An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. In: International Conference on Learning Representations (ICLR) (2021)

    2021

  40. [40]

    IEEE Robotics & Automation Magazine (2006)

    Durrant-Whyte, H., Bailey, T.: Simultaneous Localization and Mapping: Part I. IEEE Robotics & Automation Magazine (2006)

    2006

  41. [41]

    Computer (1989)

    Elfes, A.: Using Occupancy Grids for Mobile Robot Perception and Navigation. Computer (1989)

    1989

  42. [42]

    In: Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2024)

    Fang, X., Garrett, C.R., Eppner, C., Lozano-Pérez, T., Kaelbling, L.P., Fox, D.: DiMSam: Diffusion Models as Samplers for Task and Motion Planning under Partial Observability. In: Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2024)

    2024

  43. [43]

    Com- munications of the ACM (1981)

    Fischler, M.A., Bolles, R.C.: Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography. Com- munications of the ACM (1981)

    1981

  44. [44]

    In: Proceedings of the International Con- ference on Machine Learning (ICML) (2016)

    Gal, Y., Ghahramani, Z.: Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning. In: Proceedings of the International Con- ference on Machine Learning (ICML) (2016)

    2016

  45. [45]

    Foundations and Trends® in Robotics (2020)

    Garg,S.,Sünderhauf,N.,Dayoub,F.,Morrison,D.,Cosgun,A.,Carneiro,G.,Wu, Q., Chin, T.J., Reid, I., Gould, S., Corke, P., Milford, M.: Semantics for Robotic Mapping, Perception and Interaction: A Survey. Foundations and Trends® in Robotics (2020)

    2020

  46. [46]

    Journal of the American Statistical Association (2007)

    Gneiting, T., Raftery, A.E.: Strictly Proper Scoring Rules, Prediction, and Esti- mation. Journal of the American Statistical Association (2007)

    2007

  47. [47]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2014) 18 Bhattacharjee et al

    Goodfellow, I.J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., Bengio, Y.: Generative Adversarial Nets. In: Advances in Neural Information Processing Systems (NeurIPS) (2014) 18 Bhattacharjee et al

    2014

  48. [48]

    In: Conference on Robot Learning (CoRL) (2020)

    Han, Y., Banfi, J., Campbell, M.: Planning Paths Through Unknown Space by Imagining What Lies Therein. In: Conference on Robot Learning (CoRL) (2020)

    2020

  49. [49]

    Cam- bridge University Press (2004)

    Hartley, R., Zisserman, A.: Multiple View Geometry in Computer Vision. Cam- bridge University Press (2004)

    2004

  50. [50]

    Zeitschrift für Mathematik und Physik (1900)

    Heun, K.: Neue methode zur approximativen lösung der differentialgleichungen einer unabhängigen veränderlichen. Zeitschrift für Mathematik und Physik (1900)

    1900

  51. [51]

    In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2025)

    Ho, C., Kim, S., Moon, B., Parandekar, A., Harutyunyan, N., Wang, C., Sycara, K., Best, G., Scherer, S.: MapEx: Indoor structure exploration with probabilis- tic information gain from global map predictions. In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2025)

    2025

  52. [52]

    In: Ad- vances in Neural Information Processing Systems (NeurIPS) (2020)

    Ho, J., Jain, A., Abbeel, P.: Denoising Diffusion Probabilistic Models. In: Ad- vances in Neural Information Processing Systems (NeurIPS) (2020)

    2020

  53. [53]

    In: NeurIPS Workshop on Diffusion Models (2022)

    Ho, J., Salimans, T.: Classifier-Free Diffusion Guidance. In: NeurIPS Workshop on Diffusion Models (2022)

    2022

  54. [54]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023)

    Hollein, L., Cao, A., Owens, A., Johnson, J., Niessner, M.: Text2Room: Extract- ing Textured 3D Meshes from 2D Text-to-Image Models. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023)

    2023

  55. [55]

    In: Proceedings of the International Conference on Machine Learning (ICML) (2023)

    Hoogeboom, E., Heek, J., Salimans, T.: Simple Diffusion: End-to-End Diffusion for High Resolution Images. In: Proceedings of the International Conference on Machine Learning (ICML) (2023)

    2023

  56. [56]

    Hoogeboom, E., Nielsen, D., Jaini, P., Forré, P., Welling, M.: Argmax Flows and MultinomialDiffusion:LearningCategoricalDistributions.In:AdvancesinNeural Information Processing Systems (NeurIPS) (2021)

    2021

  57. [57]

    Autonomous Robots (2013)

    Hornung, A., Wurm, K.M., Bennewitz, M., Stachniss, C., Burgard, W.: OctoMap: An efficient probabilistic 3D mapping framework based on octrees. Autonomous Robots (2013)

    2013

  58. [58]

    In: International Conference on 3D Vision (3DV) (2016)

    Hua, B.S., Pham, Q.H., Nguyen, D.T., Tran, M.K., Yu, L.F., Yeung, S.K.: Sce- neNN: A scene meshes dataset with annotations. In: International Conference on 3D Vision (3DV) (2016)

    2016

  59. [59]

    Bulletin de la Société Vaudoise des Sciences Naturelles (1901)

    Jaccard, P.: Étude comparative de la distribution florale dans une portion des Alpes et des Jura. Bulletin de la Société Vaudoise des Sciences Naturelles (1901)

    1901

  60. [60]

    In: Proceedings of the European Conference on Computer Vision (ECCV) (2016)

    Johnson, J., Alahi, A., Fei-Fei, L.: Perceptual Losses for Real-Time Style Transfer and Super-Resolution. In: Proceedings of the European Conference on Computer Vision (ECCV) (2016)

    2016

  61. [61]

    Artificial Intelligence (1998)

    Kaelbling,L.P.,Littman,M.L.,Cassandra,A.R.:PlanningandActinginPartially Observable Stochastic Domains. Artificial Intelligence (1998)

    1998

  62. [62]

    In: Advances in Neural Information Process- ing Systems (NeurIPS) (2022)

    Karras, T., Aittala, M., Aila, T., Laine, S.: Elucidating the Design Space of Diffusion-Based Generative Models. In: Advances in Neural Information Process- ing Systems (NeurIPS) (2022)

    2022

  63. [63]

    Advanced Robotics (2022)

    Katsumata, Y.,Kanechika, A.,Taniguchi,A., Hafi,L.E.,Hagiwara,Y.,Taniguchi, T.: Map Completion from Partial Observation using the Global Structure of Mul- tiple Environmental Maps. Advanced Robotics (2022)

    2022

  64. [64]

    In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2019)

    Katyal, K., Popek, K., Paxton, C., Burlina, P., Hager, G.D.: Uncertainty-Aware Occupancy Map Prediction Using Generative Networks for Robot Navigation. In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2019)

    2019

  65. [65]

    ACM Transactions on Graphics (2023)

    Kerbl, B., Kopanas, G., Leimkühler, T., Drettakis, G.: 3D Gaussian Splatting for Real-Time Radiance Field Rendering. ACM Transactions on Graphics (2023)

    2023

  66. [66]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2018)

    Kohl, S.A.A., Romera-Paredes, B., Meyer, C., Fauw, J.D., Ledsam, J.R., Maier- Hein, K.H., Eslami, S.M.A., Rezende, D.J., Ronneberger, O.: A Probabilistic U- Generative Single-View Floormap Completion 19 Net for Segmentation of Ambiguous Images. In: Advances in Neural Information Processing Systems (NeurIPS) (2018)

    2018

  67. [67]

    Artificial Intelligence (2000)

    Kuipers, B.: The Spatial Semantic Hierarchy. Artificial Intelligence (2000)

    2000

  68. [68]

    Annual Review of Control, Robotics, and Autonomous Systems (2022)

    Kurniawati, H.: Partially Observable Markov Decision Processes and Robotics. Annual Review of Control, Robotics, and Autonomous Systems (2022)

    2022

  69. [69]

    Cambridge University Press (2006)

    LaValle, S.M.: Planning Algorithms. Cambridge University Press (2006)

    2006

  70. [70]

    In: International Conference on Learning Rep- resentations (ICLR) (2026)

    Li, H., Zheng, W., He, J., Liu, Y., Lin, X., Yang, X., Chen, Y.C., Guo, C.: DA2: Depth anything in any direction. In: International Conference on Learning Rep- resentations (ICLR) (2026)

    2026

  71. [71]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022)

    Li, J., Chen, C., Xiong, Z.: Contextual outpainting with object-level contrastive learning. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022)

    2022

  72. [72]

    Machine Intelligence Research (2026)

    Li, X., Qiu, H., Wang, L., Zhang, H., Qi, C., Han, L., Xiong, H., Li, H.: Challenges and Trends in Egocentric Vision: A Survey. Machine Intelligence Research (2026)

    2026

  73. [73]

    In: Proceedings of the European Conference on Computer Vision (ECCV) (2022)

    Li, Z., Wang, W., Li, H., Xie, E., Sima, C., Lu, T., Qiao, Y., Dai, J.: BEV- Former: Learning Bird’s-Eye-View representation from multi-camera images via spatiotemporal transformers. In: Proceedings of the European Conference on Computer Vision (ECCV) (2022)

    2022

  74. [74]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    Ling, L., Sheng, Y., Tu, Z., Zhao, W., Xin, C., Wan, K., Yu, L., Guo, Q., Yu, Z., Lu, Y., Li, X., Sun, X., Ashok, R., Mukherjee, A., Kang, H., Kong, X., Hua, G., Zhang, T., Benes, B., Bera, A.: DL3DV-10K: A large-scale scene dataset for deep learning-based 3D vision. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CV...

    2024

  75. [75]

    Journal of Mathematical Chemistry (1999)

    Lipkus, A.H.: A proof of the triangle inequality for the Tanimoto distance. Journal of Mathematical Chemistry (1999)

    1999

  76. [76]

    In: International Conference on Learning Representations (ICLR) (2023)

    Lipman, Y., Chen, R.T.Q., Ben-Hamu, H., Nickel, M., Le, M.: Flow Matching for Generative Modeling. In: International Conference on Learning Representations (ICLR) (2023)

    2023

  77. [77]

    arXiv preprint arXiv:2412.06264 (2024)

    Lipman, Y., Havasi, M., Holderrieth, P., Shaul, N., Le, M., Karrer, B., Chen, R.T.Q., Lopez-Paz, D., Ben-Hamu, H., Gat, I.: Flow Matching Guide and Code. arXiv preprint arXiv:2412.06264 (2024)

    Pith/arXiv arXiv 2024

  78. [78]

    In: Proceedings of the European Conference on Com- puter Vision (ECCV) (2018)

    Liu, C., Wu, J., Furukawa, Y.: FloorNet: A unified framework for floorplan recon- struction from 3D scans. In: Proceedings of the European Conference on Com- puter Vision (ECCV) (2018)

    2018

  79. [79]

    In: Proceedings of the European Conference on Computer Vision (ECCV) (2018)

    Liu, G., Reda, F.A., Shih, K.J., Wang, T.C., Tao, A., Catanzaro, B.: Image In- painting for Irregular Holes Using Partial Convolutions. In: Proceedings of the European Conference on Computer Vision (ECCV) (2018)

    2018

  80. [80]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    Liu, M., Xu, C., Jin, H., Gao, L., Tao, X., Li, Y.: One-2-3-45++: Fast Single Image to 3D Objects with Consistent Multi-View Generation and 3D Diffusion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    2024

Showing first 80 references.