jmkey/spatial_mosaic_vqa

SpatialMosaic: A Multi-View VLM Dataset for Partial Visibility

A multi-view visual question answering dataset for evaluating spatial reasoning under partial visibility, occlusion, and low-overlap views.

SpatialMosaic pairs indoor ScanNet++ and outdoor Waymo scene references with multi-frame VQA annotations. Questions require models to combine fragmented evidence across 2-5 views, rather than answering from a single image. The tasks cover object counting, object presence, localization, best-view selection, object-object spatial relations, and view-specific position reasoning.

This Hugging Face repository contains annotation files only. It does not redistribute ScanNet++, Waymo, image, video, depth, or scene data. Users must obtain any required source images from the official dataset providers and comply with their licenses, access requirements, and usage terms.

SpatialMosaic at a glance:

• 12,060 examples in the current tiny-test config
• 12 task types, with 1,005 examples per task type
• Multi-view inputs with 2-5 referenced frames per example
• Current tiny test split references 15 indoor ScanNet++ scenes
• Evaluation format: multiple-choice answer matching via mc_answer
• Designed for partial visibility, occlusion-aware reasoning, and low-overlap multi-view observations

Available Files and Variants

This card is configured for the uploaded tiny-test split. The SpatialMosaic project README also describes a larger annotation archive with indoor and outdoor train/test files.

Variant / split	Source reference	File	Examples / entries	Status
tiny-test / test	ScanNet++	merged_tiny_testset.json	12,060	Current config
train	ScanNet++	spatial_mosaic_vqa/train/spatialmosaic_indoor_train.json	234,979	Full project archive
train	Waymo	spatial_mosaic_vqa/train/spatialmosaic_outdoor_train.json	28,845	Full project archive
test	ScanNet++	spatial_mosaic_vqa/test/spatialmosaic_indoor_test.json	126,795	Full project archive
test	Waymo	spatial_mosaic_vqa/test/spatialmosaic_outdoor_test.json	12,688	Full project archive

If you publish the full archive as a loadable datasets config, update the front-matter configs and size_categories accordingly.

Loading

from datasets import load_dataset

ds = load_dataset("jmkey/spatial_mosaic_vqa", "tiny-test")
print(ds["test"][0])

For direct local use of the raw JSON:

import json

with open("merged_tiny_testset.json", "r") as f:
    rows = json.load(f)

Intended Use

SpatialMosaic is intended for non-commercial research on:

• multi-view visual question answering
• spatial reasoning under partial visibility
• occlusion-aware and low-overlap view understanding
• VLM evaluation and instruction-tuning on multi-frame inputs

Out-of-scope uses include commercial use under this annotation license, redistribution of restricted source datasets, and treating these annotations as a substitute for the original ScanNet++ or Waymo data.

Citation

If you use SpatialMosaic, cite the paper:

@article{lee2025spatialmosaic,
  title={SpatialMosaic: A Multiview VLM Dataset for Partial Visibility},
  author={Lee, Kanghee and Lee, Injae and Kwak, Minseok and Hong, Jungi and Ryu, Kwonyoung and Park, Jaesik},
  journal={arXiv preprint arXiv:2512.23365},
  year={2025}
}

License

The released VQA annotations are licensed under CC BY-NC 4.0. This license applies only to the annotations. Source images, videos, depth, and scene data remain governed by the original ScanNet++ and Waymo licenses and terms.