TartanAir-Absolute-Camera

# TartanAir-Absolute-Camera  Per-frame **camera parameter** annotations for the **TartanAir** dataset (**18 environments scenes**; **306,616** valid per-frame annotations), captioned by the [**Puffin-World**](https://github.com/KangLiao929/Puffin) model. The collage above visualizes the camera maps on sample images — each pair shows the **up field** (green arrows: the projected gravity-up direction) and the **latitude field** (colored contours: angle above/below the horizon). ## Format **One `.zip` per environment** — `<environment>.zip` (e.g. `abandonedfactory.zip`), 18 environments; each archive holds all its Easy/Hard trajectories. Each zip unpacks to the original per-frame layout: ``` Easy/P000/000000_camera.json Hard/P001/000012_camera.json ... ``` Each JSON holds the predicted monocular camera parameters: | field | meaning | unit | |-------|---------|------| | `roll` | camera roll | radians | | `pitch` | camera pitch | radians | | `vfov` | vertical field of view | radians | | `k1` | radial distortion coefficient | – | | `parse_ok` | whether the model output parsed within valid ranges | bool | Example: ```json {"roll": -0.0014, "pitch": -0.0764, "vfov": 1.2465, "k1": 0.0, "parse_ok": true} ``` ## Camera Parameter Distributions Histograms of the predicted roll / pitch / vertical-FoV over all frames (proportion of valid samples per 10° bin; `parse_ok=False` samples excluded).  | split | roll μ / med / σ | pitch μ / med / σ | FoV μ / med / σ | |-------|------------------|-------------------|-----------------| | all (307K frames) | -0.1° / -0.1° / 15.1° | -5.3° / -4.4° / 23.8° | 68.9° / 71.4° / 7.3° | **Reading the distributions** - **Roll** is spread around 0° (σ ≈ 15.1°) — captures are shot close to level. - **Pitch** carries a slight downward bias (median ≈ -4.4°, mean ≈ -5.3°) with σ ≈ 7–24° spread. - **FoV** concentrates around **64–79°** (median ≈ 71.4°). If you'd like a dataset with a more diverse and uniform distribution of camera parameters, please refer to our [Puffin-4M](https://huggingface.co/datasets/KangLiao/Puffin-4M) and [Puffin-16M](https://huggingface.co/datasets/KangLiao/Puffin-16M) datasets. ### Dataset Download You can download the entire dataset using the following command: ```bash hf download KangLiao/TartanAir-Absolute-Camera --repo-type dataset ``` ### Caption Pipeline Beyond this captioned dataset, we also release **a complete captioning pipeline** for annotating camera parameters for arbitrary datasets, analyzing camera parameter distributions, and visualizing the corresponding camera maps. The pipeline is available in our [GitHub repository](https://github.com/KangLiao929/Puffin). ### Citation If you find the captioned dataset useful for your research or applications, please cite our paper using the following BibTeX: ```bibtex @article{liao2025puffin, title={Thinking with Camera: A Unified Multimodal Model for Camera-Centric Understanding and Generation}, author={Liao, Kang and Wu, Size and Wu, Zhonghua and Jin, Linyi and Wang, Chao and Wang, Yikai and Wang, Fei and Li, Wei and Loy, Chen Change}, journal={arXiv preprint arXiv:2510.08673}, year={2025} } ```