| --- |
| annotations_creators: [] |
| language: [] |
| language_creators: [] |
| license: |
| - cc0-1.0 |
| multilinguality: [] |
| pretty_name: Anime Segmentation |
| size_categories: |
| - 10K<n<100K |
| source_datasets: |
| - original |
| tags: [] |
| task_categories: |
| - image-segmentation |
| task_ids: |
| - semantic-segmentation |
| --- |
| |
| ## Dataset Description |
|
|
| A segmentation dataset for anime character |
|
|
| My project: [anime-segmentation](https://github.com/SkyTNT/anime-segmentation) |
|
|
| ### Dataset Summary |
|
|
| | Dir | Description | Format | Images | |
| | ---- | ---- | ---- | ---- | |
| | bg | background images | jpg | 8057 | |
| | fg | foreground images, transparent background | png | 11802 | |
| | imgs | real images with background and foreground| jpg | 1111 | |
| | masks| labels for imgs | jpg | 1111 | |
|
|
| Total size: 18GB |
|
|
| ### Collection Method |
|
|
| Collect background from [character_bg_seg_data](https://github.com/ShuhongChen/bizarre-pose-estimator#download) |
|
|
| Collect foreground from danbooru website. |
|
|
| Collect imgs and masks from [AniSeg](https://github.com/jerryli27/AniSeg#about-the-models) and danbooru website. |
|
|
| I use [Real-ESRGAN](https://github.com/xinntao/Real-ESRGAN) to restore the background images. |
|
|
| I clean the dataset using [DeepDanbooru](https://github.com/KichangKim/DeepDanbooru) first then manually, to make sue all foreground is anime character. |
|
|
| ### Contributions |
|
|
| Thanks to [@SkyTNT](https://github.com/SkyTNT) for adding this dataset. |
|
|
| Thanks to [@ShuhongChen](https://github.com/ShuhongChen) for [character_bg_seg_data](https://github.com/ShuhongChen/bizarre-pose-estimator#download) |
|
|
| Thanks to [@jerryli27](https://github.com/jerryli27) for [AniSeg](https://github.com/jerryli27/AniSeg#about-the-models) |
|
|
