Megalith-10M-Camera

# Megalith-10M-Camera  Per-image camera parameter annotations for the **Megalith-10M** dataset (the image-bearing build `drawthingsai/megalith-10m`; ~9.58M Flickr photos across 959 shards), 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 `.tar` per source shard (`00000.tar` … `00958.tar`), each containing one `.json` per image (name matches the source image stem). Each JSON holds the predicted monocular camera parameters plus the source URL: | 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 | | `url` | source image URL (Flickr static) | – | Example: ```json {"roll": 0.0123, "pitch": -0.0871, "vfov": 1.0123, "k1": 0.0000, "parse_ok": true, "url": "https://farm4.staticflickr.com/3862/15041414305_58f2d67117_b.jpg"} ``` Every caption JSON carries the `url` of its source image, so each annotation is self-contained and can be linked back to the image without the source dataset. ## Camera Parameter Distributions Histograms of the predicted roll / pitch / vertical-FoV over the whole dataset (proportion of valid samples per 10° bin; `parse_ok=False` excluded).  | split | roll μ / med / σ | pitch μ / med / σ | FoV μ / med / σ | |-------|------------------|-------------------|-----------------| | all (9.38M) | 0.0° / 0.0° / 6.7° | −3.1° / −2.3° / 12.1° | 34.6° / 31.3° / 12.0° | - **Roll** is sharply peaked at 0° (web images shot upright/level). - **Pitch** is centered slightly negative (mild downward-looking tendency). - **FoV** has the widest spread of the three (σ ≈ 12°), long wide-angle tail — consistent with the diverse focal lengths of unconstrained Flickr photography. 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/Megalith-10M-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} } ```