FlatLands Dataset

FlatLands is the official dataset release for Generative Floormap Completion From a Single Egocentric View.

Dataset Card Status

This Hugging Face repository contains:

• README.md: this dataset card.
• FlatLands_final_dataset.zip: the validated dataset archive.
• LICENSE: FlatLands release notice.
• LICENSES.md: upstream source dataset license and terms pointers.
• train.parquet, validation.parquet, test.parquet: small Hub-native preview splits used by the Hugging Face Dataset Viewer.

The Dataset Viewer shows a small preview subset with one preview_image mosaic per row, combining the observed floor, target floor map, unobserved mask, and epistemic mask. The complete 270,575-observation dataset is in FlatLands_final_dataset.zip.

Project code, construction details, provenance documentation, and paper artifacts are maintained in the main GitHub repository: https://github.com/1ssb/Flat_Lands/

Dataset Summary

• Observations: 270,575
• Real metric indoor scenes: 17,656
• Source datasets: 6
• Splits: train, validation, test
• Observation packet: observed_floor.png, floor_map.png, unobserved.png, epistemic_mask.png, metadata.json
• Filtering threshold: conditional_signal_ratio >= 0.10

Split Counts

Split	Observations
Train	215,342
Validation	26,890
Test	28,343
Total	270,575

Download

Download the archive from Hugging Face:

hf download Rudra1ssb/FlatLands FlatLands_final_dataset.zip --repo-type dataset
unzip FlatLands_final_dataset.zip -d FlatLands

The Dataset Viewer preview can also be loaded with:

from datasets import load_dataset

preview = load_dataset("Rudra1ssb/FlatLands")
print(preview)

Integrity

• Archive: FlatLands_final_dataset.zip
• Size: 2,054,773,316 bytes
• SHA-256: e4f2e5c7c54f7ba62ea696fb103fb5d3794f30f5a2e63715773e59d6a9f1d26f

Verify after download:

sha256sum FlatLands_final_dataset.zip

Dataset Structure

After extraction, the archive has this structure:

FlatLands/
├── train/
│   └── obs_*/
├── val/
│   └── obs_*/
└── test/
    └── obs_*/

Each observation directory contains exactly:

observed_floor.png
floor_map.png
unobserved.png
epistemic_mask.png
metadata.json

The files in each observation packet are:

File	Meaning
observed_floor.png	Observed floor conditioning map.
floor_map.png	Complete floor-map target.
unobserved.png	Unobserved completion-region mask.
epistemic_mask.png	Valid support / epistemic mask.
metadata.json	Observation, scene, crop, and provenance metadata.

Quick Inspection

from pathlib import Path
import json
from PIL import Image

obs = Path("FlatLands/train/obs_084303")

metadata = json.loads((obs / "metadata.json").read_text())
observed = Image.open(obs / "observed_floor.png")
target = Image.open(obs / "floor_map.png")

print(metadata["scene"]["dataset"], metadata["scene"]["scene_id"])
print(observed.size, target.size)

Citation

@inproceedings{bhattacharjee2026flatlands,
  title     = {FlatLands: Generative Floormap Completion From a Single Egocentric View},
  author    = {Bhattacharjee, Subhransu S. and Campbell, Dylan and Shome, Rahul},
  booktitle = {European Conference on Computer Vision (ECCV)},
  year      = {2026}
}

Please also consider citing the underlying datasets as used:

@inproceedings{mp3d,
  title     = {{Matterport3D}: Learning from {RGB-D} Data in Indoor Environments},
  author    = {Chang, Angel X. and Dai, Angela and Funkhouser, Thomas and Halber, Maciej and Nie{\ss}ner, Matthias and Savva, Manolis and Song, Shuran and Zeng, Andy and Zhang, Yinda},
  booktitle = {International Conference on 3D Vision (3DV)},
  pages     = {667--676},
  year      = {2017}
}

@inproceedings{dai2017scannet,
  title     = {{ScanNet}: Richly-Annotated 3D Reconstructions of Indoor Scenes},
  author    = {Dai, Angela and Chang, Angel X. and Savva, Manolis and Halber, Maciej and Funkhouser, Thomas and Nie{\ss}ner, Matthias},
  booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  pages     = {2432--2443},
  year      = {2017}
}

@inproceedings{yeshwanth2023scannetpp,
  title     = {{ScanNet++}: A High-Fidelity Dataset of 3D Indoor Scenes},
  author    = {Yeshwanth, Chandan and Liu, Yueh-Cheng and Nie{\ss}ner, Matthias and Dai, Angela},
  booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
  year      = {2023}
}

@inproceedings{baruch2021arkitscenes,
  title     = {{ARKitScenes}: A Diverse Real-World Dataset for 3D Indoor Scene Understanding Using Mobile {RGB-D} Data},
  author    = {Baruch, Gilad and Chen, Zhuoyuan and Dehghan, Afshin and Dimry, Tal and Feigin, Yuri and Fu, Peter and Gebauer, Thomas and Joffe, Brandon and Kurz, Daniel and Schwartz, Arik and Shulman, Elad},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS) Datasets and Benchmarks Track},
  year      = {2021}
}

@inproceedings{wald2019rio,
  title     = {{RIO}: 3D Object Instance Re-Localization in Changing Indoor Environments},
  author    = {Wald, Johanna and Avetisyan, Armen and Navab, Nassir and Tombari, Federico and Nie{\ss}ner, Matthias},
  booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
  pages     = {7657--7666},
  year      = {2019}
}

@inproceedings{zind,
  title     = {{Zillow Indoor Dataset}: Annotated Floor Plans with 360deg Panoramas and 3D Room Layouts},
  author    = {Cruz, Steve and Hutchcroft, Will and Li, Yuguang and Khosravan, Naji and Boyadzhiev, Ivaylo and Kang, Sing Bing},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  pages     = {2133--2143},
  year      = {2021}
}