Skip to content

zhenglinpan/AnitaDataset

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 

Repository files navigation

Icon Anita Dataset - An Industrial Animation Dataset


image

Icon Introduction

This is the official repository for Anita Dataset.

Anita Dataset is a collection of 16,000+ professional hand-drawn cartoon keyframes. The dataset contains 1080P sketch, color images and composition exported from engineering files of 14 different animations(more than 200 scenes in total), following the professional standards of the modern 2D animation industry.

Sketch Color Composition

Cartoon image sequence are provided in .png format, licensed by animators under CC BY-NC-SA 4.0 and CC BY License. Here we show some examples of the dataset:

Example 1 - Sketch Example 1 - Color
Example 2 - Sketch Example 2 - Composition
Example 3 - Sketch Example 3 - Color
Example 4 - Sketch Example 4 - Composition
image

We designed this dataset to facilitate cartoon research, such as inbetweening and colorization. Researchers can use the dataset to train and evaluate their models.

For more Information, please refer to Q&A section.

Icon Download

We provide two download links for Anita Dataset:

Google Drive

百度网盘

The file size of current Anita Dataset is 7.14G, feel free to download and explore it!

Information

Preview Animation Artist(A-Z) Keyframe Sketch Clip Color Clip Composition Clip License
image Dogmatism _嘲讽怪_ 3,559 12 7 0 CC BY-NC-SA 4.0
image Mirror 剑起长歌魂断蓝桥 1,098 15 15 0 CC BY-NC-SA 4.0
image Embrace Freedom 日勿广隶 5,338 58 0 1 CC BY-NC-SA 4.0
image Zixu's 2023 Reels 秄虚 1,079 9 3 3 CC BY-NC-SA 4.0
image Big Bad Fox Christoph Lendenfeld 42 1 1 1 CC BY
image Hero Daniel Martinez Lara 1,797 8 8 8 CC BY
image Trapeze Artist Jordi Porras 261 1 1 1 CC BY
image Dance Lien-ze, Tsao 336 1 1 1 CC BY
image Explosion Nahu 34 1 1 0 CC BY
image Hope ONE_XT 1,478 69 63 0 CC BY-NC-SA 4.0
image Polaris ONE_XT 586 34 34 0 CC BY-NC-SA 4.0
image Sierra Dragon Pepe Sanchez 60 1 1 1 CC BY
image LMA test Renato Roldan 570 1 1 1 CC BY
image Pirate Sergi Miranda 618 1 1 1 CC BY

Icon File Structure

Anita_Dataset
├── Mirror
│   ├── scene_#
│   │   ├── sketch
│   │   │   ├── 0001.png
│   │   │   ├── 0002.png
│   │   │   ├── ...
│   │   ├── color
│   │   │   ├── 0001.png
│   │   │   ├── 0002.png
│   │   │   ├── ...
│   │   ├── composition
│   │   │   ├── 0001.png
│   │   │   ├── 0002.png
│   │   │   ├── ...
│   ├── scene_#
│   │   ├── ...
├── Embrace_Freedom
│   ├── scene_#
│   │   ├── sketch
│   │   │   ├── ...

Icon Q&A

Does Anita Dataset have any related publications?

No. Anita Dataset does not have an associated research paper. But researchers are welcome to use this dataset in their own studies.

What is Anita Dataset different from other cartoon datasets?

  • Hand-drawn Authenticity: Unlike AnimeRun or PBC, which are rendered from 3D character models, Anita Dataset consists of frames drawn by human hands, aligning more closely with the standards of the modern 2D animation industry.

  • High-quality Intermediate Format: Compared to Sakuga-42M Dataset, which provides large-scale but 480p rendered clips, Anita Dataset offers high-resolution intermediate images. More than 16,000 sketches and flat colors are provided in 1080p resolution.

  • Sketch-Color Alignment : Anita Dataset contains sketch-color pairs exported directly from engineering files.

Is the dataset licensed?

  • Yes, Anita Dataset is licensed. The animations included are obtained either directly from artists with permission or under distributable licenses from which they are shared.

Are there annotations provided ?

  • No. If researchers need annotations, they can create their own based on the dataset.

Are the sketches and color images paired?

  • Some of them are paired, and some are not. Researchers can pair them based on the folder name.

Will the dataset be updated in the future?

  • Yes. The dataset is an ongoing project and will be updated to include more available animations.

Icon Citation

If you find this dataset useful, please consider citing our work:

@misc{Anita2024,
  title = {Anita Dataset},
  howpublished = {https://zhenglinpan.github.io/AnitaDataset_homepage/},
  note = {Accessed: 2024-06-25}
}