ityizNola/Anime-LineArt-Dataset

---
license: cc-by-nc-4.0
task_categories:
- image-to-image
- unconditional-image-generation
language:
- en
tags:
- anime
- sketch
- lineart
- controlnet
- image-processing
- text-to-sketch
- synthetic
size_categories:
- 1K<n<10K
---

# Anime Lineart Sketch Dataset

## Dataset Summary

This dataset contains high-quality **lineart sketch images** automatically extracted from raw anime images using the [`LineartAnimeDetector`](https://github.com/lllyasviel/ControlNet-v1-1-nightly) model from ControlNet Annotators (`lllyasviel/Annotators`).

It is designed to support research and development in:
- Anime-style sketch generation
- Text-to-sketch pipelines
- ControlNet conditioning
- Sketch-to-image and image-to-sketch translation

The raw source images are sourced from the [Anime Images Dataset](https://www.kaggle.com/datasets/diraizel/anime-images-dataset) on Kaggle.

---

## Dataset Structure

Sketches are organized into subdirectories mirroring the original anime category structure:

```
sketches/
├── category_1/
│ ├── image_001.png
│ ├── image_002.png
│ └── ...
├── category_2/
│ └── ...
└── category_n/
└── ...
```

All images are saved as **PNG** files at **512×512** resolution.

---

## Dataset Details

| Property | Value |
|---|---|
| **Image format** | PNG |
| **Resolution** | 512 × 512 px |
| **Color mode** | Grayscale (lineart) |
| **Source dataset** | `diraizel/anime-images-dataset` (Kaggle) |
| **Extraction model** | `LineartAnimeDetector` — `lllyasviel/Annotators` |
| **Max images per category** | 15 |
| **License** | CC BY-NC 4.0 |

---

## Generation Process

Sketches were generated using an automated pipeline:

1. **Download** — Raw anime images fetched from Kaggle via `kagglehub`
2. **Extract** — Each image passed through `LineartAnimeDetector` at 512px resolution
3. **Store** — Results saved as PNG, preserving the original category subfolder structure

The extraction script is available at:
🔗 [github.com/iCogLabs/text-to-sketch](https://github.com/iCogLabs/text-to-sketch) *(placeholder — update with actual repo URL)*

---

## Intended Use

### ✅ Intended Uses
- Training sketch-conditioned diffusion models (e.g., ControlNet)
- Building text-to-sketch or sketch-to-image datasets
- Anime style transfer and lineart research

### ❌ Out-of-Scope Uses
- Commercial use without explicit permission (see license)
- Generating content that infringes copyright of original anime works

---

## Source Data

The raw images are sourced from:

> **Anime Images Dataset** by [diraizel](https://www.kaggle.com/datasets/diraizel/anime-images-dataset) on Kaggle.
> Original images are not redistributed in this dataset — only the extracted lineart representations are included.

---

## Citation

If you use this dataset in your research or project, please cite:

```bibtex
@dataset{icoglabs_anime_sketches_2025,
author = {iCog Labs},
title = {Anime Lineart Sketch Dataset},
year = {2025},
publisher = {Hugging Face},
url = {https://huggingface.co/datasets/ityizNola/Anime-LineArt-Dataset}
}
```

---

## Contact

Maintained by **iCog Labs** — AI Video Generation Research Team.
For questions or contributions, open an issue on the source repository.