Datasetpublic

nviolante/dot3d

Lighting-Consistent Object Transfer Across Radiance Fields — 3D Scenes Dataset This dataset accompanies the paper Lighting-Consistent Object Transfer Across Radiance Fields (Eurographics Symposium on Rendering 2026). Paper: https://inria.hal.science/hal-05657202v1Code: github.com/graphdeco-inria/dot3dModel weights: nviolante/dot3d Dataset structure dot3d_data/ ├── data_3d/ │ ├── real/insertions/ # 7 real-world insertion scenes │ └── synthetic/ │… See the full description on the dataset page: https://huggingface.co/datasets/nviolante/dot3d.

sourceHugging Faceotherupdated 3mo agoView on Hugging Face

0likes889downloads

Dataset Card

Lighting-Consistent Object Transfer Across Radiance Fields — 3D Scenes Dataset

This dataset accompanies the paper Lighting-Consistent Object Transfer Across Radiance Fields (Eurographics Symposium on Rendering 2026).

Paper: https://inria.hal.science/hal-05657202v1 Code: github.com/graphdeco-inria/dot3d Model weights: nviolante/dot3d

Dataset structure

dot3d_data/
├── data_3d/
│   ├── real/insertions/        # 7 real-world insertion scenes
│   └── synthetic/
│       ├── insertions/         # 9 synthetic insertion scenes
│       └── scenes/             # Raw background and object captures
│           ├── backgrounds/
│           └── objects/
├── data_rotation/              # 3 scenes × 3 rotation variants (9 total)
├── results_3d/
│   ├── real/                   # Pre-computed results on real scenes
│   └── synthetic/              # Pre-computed results on synthetic scenes
└── results_rotation/           # Pre-computed results for the rotation experiment

Scenes

3D — Real insertions (`data_3d/real/insertions`)

Scene
deskinlondon_bedroom
locomotiveintrain
plantinbouquet
sedaninc0e46b
truckingarden
508850in4cea29
d11d95in83d5f2

3D — Synthetic insertions (`data_3d/synthetic/insertions`)

Scene
bathtub57einblend57eadb
bedbg3inblendadfc14
carriageinblend_85a1a6
helicopterinblend_808c29
sinkinblend_92e6c7
sofabg3inblend0ab265
sofablend057inblend_057cff
table7eainblend7ea0f8
whitetruckinblendb433c8

Rotation experiment (`data_rotation`)

Three scenes evaluated across three object rotations (rot0, rot1, rot2): bathtub_57e, sink, white_truck.

Data format

Each scene is a COLMAP capture. Insertion scenes include a masks/ folder with a binary mask per frame:

<scene>/
├── images/
├── masks/        # binary PNG masks, one per frame (insertion scenes only)
└── sparse/
    └── 0/
        ├── cameras.bin
        ├── images.bin
        └── points3D.bin

Download

bash

# 3D scenes
huggingface-cli download nviolante/dot3d_data --repo-type dataset --include "data_3d/*" --local-dir data_3d

# Rotation experiment data
huggingface-cli download nviolante/dot3d_data --repo-type dataset --include "data_rotation/*" --local-dir data_rotation

# Pre-computed results
huggingface-cli download nviolante/dot3d_data --repo-type dataset --include "results_3d/*" --local-dir results_3d
huggingface-cli download nviolante/dot3d_data --repo-type dataset --include "results_rotation/*" --local-dir results_rotation

License

This dataset is released under the same non-commercial research license as the code. See LICENSE.

Citation

bibtex

@Article{violante2026dot3d,
      author       = {Violante, Nicolás and Kopanas, George and Franke, Linus and Philip, Julien and Drettakis, George},
      title        = {Lighting-Consistent Object Transfer Across Radiance Fields},
      journal      = {Computer Graphics Forum (Proceedings of the Eurographics Symposium on Rendering)},
      year         = {2026},
      volume       = 45,
      number       = 4
}