---
pretty_name: SceneFly
license: apache-2.0
language:
  - en
  - zh
tags:
  - video
  - computer-vision
  - world-model
  - multimodal
  - camera-pose
---

# CaR

## Compression and Retrieval: Implicit Memory Retrieval for Video World Models

<a href="https://pzzz-cv.github.io/">Zhan Peng</a><sup>1,2</sup>,
Jie Ma<sup>2</sup>,
Huiqiang Sun<sup>1</sup>,
Chong Gao<sup>2,3</sup>,
Zhijie Xue<sup>1</sup>,
Zhiyu Pan<sup>1</sup>,
Zhiguo Cao<sup>1*</sup>,
Jun Liang<sup>2*</sup>,
Jing Li<sup>2</sup>

<sup>1</sup>Huazhong University of Science and Technology &nbsp;
<sup>2</sup>HUJING Digital Media & Entertainment Group &nbsp;
<sup>3</sup>Sun Yat-sen University

<sup>*</sup>Corresponding author

[![Paper](https://img.shields.io/badge/Paper-arXiv-red)](https://arxiv.org/abs/2606.23105)
[![Project Page](https://img.shields.io/badge/Project-Page-blue)](https://orange-3dv-team.github.io/CaR/)
[![Dataset](https://img.shields.io/badge/Dataset-SceneFly-yellow)](https://huggingface.co/datasets/Orange-3DV-Team/SceneFly)

## SceneFly

`SceneFly` is a curated video dataset organized by synthetic 3D scenes. Each selected video contains the source video, camera annotations, text prompt, and sample-level context/ground-truth metadata.

This release contains a subset of the SceneFly dataset filter, including 58 scenes, 462 videos, and 45,173 sample annotations. The complete dataset will be provided in a future release.

## Usage

The dataset is provided as a split `tar.gz` stream. Merge the split parts and extract the dataset with:

```bash
cat SceneFly_* | tar -xzvf -
```

The split parts are named:

```text
SceneFly_aa
SceneFly_ab
SceneFly_ac
...
```

After extraction, the dataset root will be:

```text
SceneFly/
```

A checksum manifest is provided in:

```text
manifest.csv
```

It records each split part name, part size, source dataset size, and SHA256 checksum.

## Dataset Structure

```text
SceneFly/
├── metadata.csv
├── AlbertMansion/
│   ├── 0/
│   │   ├── video.mp4
│   │   ├── camera.json
│   │   ├── prompt.txt
│   │   └── samples/
│   │       ├── sample000.json
│   │       ├── sample001.json
│   │       └── ...
│   ├── 1/
│   │   └── ...
│   └── ...
├── AsianArchitecture/
│   └── ...
└── ...
```

## File Description

- `video.mp4`: rendered scene video.
- `camera.json`: camera trajectory and intrinsic annotation for the video.
- `prompt.txt`: text description/prompt of the scene.
- `samples/sampleXXX.json`: sample-level annotation file.
- `metadata.csv`: global metadata index for all samples.
- `manifest.csv`: archive-level checksum and size manifest.

## Sample Annotation Fields

Each `samples/sampleXXX.json` contains fields such as:

- `scene_name`: scene identifier.
- `video_name`: video identifier inside the scene.
- `image_width`, `image_height`: video resolution.
- `focal_length`: camera focal length.
- `context_start_segment`, `context_end_segment`: context segment range.
- `context_length_segments`: number of context segments.
- `context_segments`: list of context segment IDs.
- `context_start_frame`: starting frame of the context window.
- `context_num_frames`: number of context frames.
- `gt_segment`: ground-truth target segment.
- `gt_start_frame`: starting frame of the ground-truth segment.
- `gt_num_frames`: number of ground-truth frames.
- `overlap_score`, `containment_score`, `final_score`: sample matching scores.

## Metadata

`metadata.csv` provides one row per sample and includes the relative sample path plus the context and ground-truth fields above.

## Citation

If you find this dataset useful, please cite:

```bibtex
@article{peng2026car,
    title={Compression and Retrieval: Implicit Memory Retrieval for Video World Models},
    author={Peng, Zhan and Ma, Jie and Sun, Huiqiang and Gao, Chong and Xue, Zhijie and Pan, Zhiyu and Cao, Zhiguo and Liang, Jun and Li, Jing},
    journal={arXiv preprint arXiv:2606.23105},
    year={2026}
}
```