CyberHarem/matsubara_kanon_bangdream
收藏Hugging Face2024-01-15 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/CyberHarem/matsubara_kanon_bangdream
下载链接
链接失效反馈官方服务:
资源简介:
---
license: mit
task_categories:
- text-to-image
tags:
- art
- not-for-all-audiences
size_categories:
- n<1K
---
# Dataset of matsubara_kanon/松原花音/마츠바라카논 (BanG Dream!)
This is the dataset of matsubara_kanon/松原花音/마츠바라카논 (BanG Dream!), containing 500 images and their tags.
The core tags of this character are `blue_hair, purple_eyes, bangs, long_hair, one_side_up, ribbon, hair_ribbon, hair_ornament, light_blue_hair`, which are pruned in this dataset.
Images are crawled from many sites (e.g. danbooru, pixiv, zerochan ...), the auto-crawling system is powered by [DeepGHS Team](https://github.com/deepghs)([huggingface organization](https://huggingface.co/deepghs)).
## List of Packages
| Name | Images | Size | Download | Type | Description |
|:-----------------|---------:|:-----------|:---------------------------------------------------------------------------------------------------------------------------|:-----------|:---------------------------------------------------------------------|
| raw | 500 | 683.74 MiB | [Download](https://huggingface.co/datasets/CyberHarem/matsubara_kanon_bangdream/resolve/main/dataset-raw.zip) | Waifuc-Raw | Raw data with meta information (min edge aligned to 1400 if larger). |
| 800 | 500 | 399.39 MiB | [Download](https://huggingface.co/datasets/CyberHarem/matsubara_kanon_bangdream/resolve/main/dataset-800.zip) | IMG+TXT | dataset with the shorter side not exceeding 800 pixels. |
| stage3-p480-800 | 1237 | 872.69 MiB | [Download](https://huggingface.co/datasets/CyberHarem/matsubara_kanon_bangdream/resolve/main/dataset-stage3-p480-800.zip) | IMG+TXT | 3-stage cropped dataset with the area not less than 480x480 pixels. |
| 1200 | 500 | 610.99 MiB | [Download](https://huggingface.co/datasets/CyberHarem/matsubara_kanon_bangdream/resolve/main/dataset-1200.zip) | IMG+TXT | dataset with the shorter side not exceeding 1200 pixels. |
| stage3-p480-1200 | 1237 | 1.22 GiB | [Download](https://huggingface.co/datasets/CyberHarem/matsubara_kanon_bangdream/resolve/main/dataset-stage3-p480-1200.zip) | IMG+TXT | 3-stage cropped dataset with the area not less than 480x480 pixels. |
### Load Raw Dataset with Waifuc
We provide raw dataset (including tagged images) for [waifuc](https://deepghs.github.io/waifuc/main/tutorials/installation/index.html) loading. If you need this, just run the following code
```python
import os
import zipfile
from huggingface_hub import hf_hub_download
from waifuc.source import LocalSource
# download raw archive file
zip_file = hf_hub_download(
repo_id='CyberHarem/matsubara_kanon_bangdream',
repo_type='dataset',
filename='dataset-raw.zip',
)
# extract files to your directory
dataset_dir = 'dataset_dir'
os.makedirs(dataset_dir, exist_ok=True)
with zipfile.ZipFile(zip_file, 'r') as zf:
zf.extractall(dataset_dir)
# load the dataset with waifuc
source = LocalSource(dataset_dir)
for item in source:
print(item.image, item.meta['filename'], item.meta['tags'])
```
## List of Clusters
List of tag clustering result, maybe some outfits can be mined here.
### Raw Text Version
| # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | Tags |
|----:|----------:|:----------------------------------|:----------------------------------|:----------------------------------|:----------------------------------|:----------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| 0 | 15 |  |  |  |  |  | epaulettes, 1girl, band_uniform, shako_cap, solo, looking_at_viewer, open_mouth, green_ribbon, holding, sleeveless, wrist_cuffs, drumsticks, blush, white_skirt, :d, instrument |
| 1 | 44 |  |  |  |  |  | 1girl, looking_at_viewer, solo, blush, bare_shoulders, short_sleeves, smile, hair_flower, frilled_sleeves, simple_background, white_background, blue_shirt, dress, upper_body, collarbone, green_ribbon, open_mouth, frilled_shirt, off_shoulder |
| 2 | 13 |  |  |  |  |  | 1girl, blush, looking_at_viewer, solo, bow, short_sleeves, striped_ribbon, twintails, alternate_hairstyle, wrist_cuffs, earrings, red_ribbon, holding_gift, valentine, frilled_dress, center_frills, frilled_sleeves, gift_box, hair_flower, heart-shaped_box, heart_hair_ornament, white_background, :d, blue_ribbon, open_mouth, thighhighs |
| 3 | 11 |  |  |  |  |  | 1girl, solo, earrings, looking_at_viewer, double_bun, blush, open_mouth, :d, dress, braid, jellyfish, long_sleeves, see-through_sleeves, shirt, hand_on_own_chest, upper_body |
| 4 | 7 |  |  |  |  |  | 1girl, bow, detached_sleeves, looking_at_viewer, smile, solo, tiara, wrist_cuffs, blush, double_bun, frills, underwater, air_bubble, earrings, bun_cover, jellyfish, midriff, sidelocks, skirt, aqua_hair, bloomers, cleavage, gem, medium_breasts, open_mouth, pink_eyes, pink_ribbon |
| 5 | 35 |  |  |  |  |  | hanasakigawa_school_uniform, long_sleeves, 1girl, sailor_dress, blush, solo, brown_dress, red_ribbon, looking_at_viewer, neck_ribbon, smile, green_ribbon, open_mouth, double-breasted, upper_body, white_background, flower, simple_background, white_sailor_collar |
| 6 | 7 |  |  |  |  |  | 1girl, bow, hair_flower, hairband, solo, white_flower, alternate_hairstyle, blush, floral_print, hair_over_shoulder, long_sleeves, ribbon_braid, single_braid, strawberry_hair_ornament, looking_at_viewer, print_dress, striped_ribbon, upper_body, white_gloves, :d, blue_dress, flower_earrings, open_mouth, pink_ribbon, upper_teeth_only, white_background, frilled_dress, frilled_shirt_collar, frilled_sleeves, red_ribbon |
| 7 | 7 |  |  |  |  |  | 1girl, blush, hair_flower, long_sleeves, looking_at_viewer, solo, wide_sleeves, green_kimono, white_background, frilled_sleeves, maid_apron, white_flower, cosplay, frilled_apron, open_mouth, pom_pom_(clothes), wa_maid, white_apron, :d, green_bow, green_ribbon, hair_between_eyes, simple_background, striped_bow, upper_body |
| 8 | 6 |  |  |  |  |  | 1girl, frills, looking_at_viewer, solo, blush, hat_bow, hat_ribbon, rabbit_ears, top_hat, chain, open_mouth, skirt, sleeveless, upper_body, :d, black_headwear, black_ribbon, blue_bow, heart_earrings, neck_ribbon, striped, watch, white_background, wrist_cuffs |
| 9 | 10 |  |  |  |  |  | 2girls, blush, open_mouth, yuri, white_background, closed_eyes, blonde_hair |
| 10 | 13 |  |  |  |  |  | 1girl, day, outdoors, solo, cloud, looking_at_viewer, bikini, blue_sky, blush, collarbone, open_mouth, cleavage, navel, :d, frills, medium_breasts, ocean, ball, holding, jewelry, standing |
| 11 | 6 |  |  |  |  |  | 1girl, blush, hetero, solo_focus, 1boy, collarbone, large_breasts, looking_at_viewer, nipples, open_mouth, penis, clothing_aside, cum_in_pussy, sex, spread_legs, sweat, vaginal, bar_censor, bikini, girl_on_top, heavy_breathing, medium_breasts, mosaic_censoring, navel, sidelocks, straddling, tears |
| 12 | 6 |  |  |  |  |  | 1girl, hetero, nipples, open_mouth, penis, sex, solo_focus, heart-shaped_pupils, medium_breasts, navel, pussy, spread_legs, 1boy, completely_nude, cum, mosaic_censoring, nose_blush, sweat, testicles, vaginal |
### Table Version
| # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | epaulettes | 1girl | band_uniform | shako_cap | solo | looking_at_viewer | open_mouth | green_ribbon | holding | sleeveless | wrist_cuffs | drumsticks | blush | white_skirt | :d | instrument | bare_shoulders | short_sleeves | smile | hair_flower | frilled_sleeves | simple_background | white_background | blue_shirt | dress | upper_body | collarbone | frilled_shirt | off_shoulder | bow | striped_ribbon | twintails | alternate_hairstyle | earrings | red_ribbon | holding_gift | valentine | frilled_dress | center_frills | gift_box | heart-shaped_box | heart_hair_ornament | blue_ribbon | thighhighs | double_bun | braid | jellyfish | long_sleeves | see-through_sleeves | shirt | hand_on_own_chest | detached_sleeves | tiara | frills | underwater | air_bubble | bun_cover | midriff | sidelocks | skirt | aqua_hair | bloomers | cleavage | gem | medium_breasts | pink_eyes | pink_ribbon | hanasakigawa_school_uniform | sailor_dress | brown_dress | neck_ribbon | double-breasted | flower | white_sailor_collar | hairband | white_flower | floral_print | hair_over_shoulder | ribbon_braid | single_braid | strawberry_hair_ornament | print_dress | white_gloves | blue_dress | flower_earrings | upper_teeth_only | frilled_shirt_collar | wide_sleeves | green_kimono | maid_apron | cosplay | frilled_apron | pom_pom_(clothes) | wa_maid | white_apron | green_bow | hair_between_eyes | striped_bow | hat_bow | hat_ribbon | rabbit_ears | top_hat | chain | black_headwear | black_ribbon | blue_bow | heart_earrings | striped | watch | 2girls | yuri | closed_eyes | blonde_hair | day | outdoors | cloud | bikini | blue_sky | navel | ocean | ball | jewelry | standing | hetero | solo_focus | 1boy | large_breasts | nipples | penis | clothing_aside | cum_in_pussy | sex | spread_legs | sweat | vaginal | bar_censor | girl_on_top | heavy_breathing | mosaic_censoring | straddling | tears | heart-shaped_pupils | pussy | completely_nude | cum | nose_blush | testicles |
|----:|----------:|:----------------------------------|:----------------------------------|:----------------------------------|:----------------------------------|:----------------------------------|:-------------|:--------|:---------------|:------------|:-------|:--------------------|:-------------|:---------------|:----------|:-------------|:--------------|:-------------|:--------|:--------------|:-----|:-------------|:-----------------|:----------------|:--------|:--------------|:------------------|:--------------------|:-------------------|:-------------|:--------|:-------------|:-------------|:----------------|:---------------|:------|:-----------------|:------------|:----------------------|:-----------|:-------------|:---------------|:------------|:----------------|:----------------|:-----------|:-------------------|:----------------------|:--------------|:-------------|:-------------|:--------|:------------|:---------------|:----------------------|:--------|:--------------------|:-------------------|:--------|:---------|:-------------|:-------------|:------------|:----------|:------------|:--------|:------------|:-----------|:-----------|:------|:-----------------|:------------|:--------------|:------------------------------|:---------------|:--------------|:--------------|:------------------|:---------|:----------------------|:-----------|:---------------|:---------------|:---------------------|:---------------|:---------------|:---------------------------|:--------------|:---------------|:-------------|:------------------|:-------------------|:-----------------------|:---------------|:---------------|:-------------|:----------|:----------------|:--------------------|:----------|:--------------|:------------|:--------------------|:--------------|:----------|:-------------|:--------------|:----------|:--------|:-----------------|:---------------|:-----------|:-----------------|:----------|:--------|:---------|:-------|:--------------|:--------------|:------|:-----------|:--------|:---------|:-----------|:--------|:--------|:-------|:----------|:-----------|:---------|:-------------|:-------|:----------------|:----------|:--------|:-----------------|:---------------|:------|:--------------|:--------|:----------|:-------------|:--------------|:------------------|:-------------------|:-------------|:--------|:----------------------|:--------|:------------------|:------|:-------------|:------------|
| 0 | 15 |  |  |  |  |  | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 1 | 44 |  |  |  |  |  | | X | | | X | X | X | X | | | | | X | | | | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 2 | 13 |  |  |  |  |  | | X | | | X | X | X | | | | X | | X | | X | | | X | | X | X | | X | | | | | | | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 3 | 11 |  |  |  |  |  | | X | | | X | X | X | | | | | | X | | X | | | | | | | | | | X | X | | | | | | | | X | | | | | | | | | | | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 4 | 7 |  |  |  |  |  | | X | | | X | X | X | | | | X | | X | | | | | | X | | | | | | | | | | | X | | | | X | | | | | | | | | | | X | | X | | | | | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 5 | 35 |  |  |  |  |  | | X | | | X | X | X | X | | | | | X | | | | | | X | | | X | X | | | X | | | | | | | | | X | | | | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 6 | 7 |  |  |  |  |  | | X | | | X | X | X | | | | | | X | | X | | | | | X | X | | X | | | X | | | | X | X | | X | | X | | | X | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | X | | | | | | | | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 7 | 7 |  |  |  |  |  | | X | | | X | X | X | X | | | | | X | | X | | | | | X | X | X | X | | | X | | | | | | | | | | | | | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | | | | | | | | | | | | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 8 | 6 |  |  |  |  |  | | X | | | X | X | X | | | X | X | | X | | X | | | | | | | | X | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | | | | | | X | | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 9 | 10 |  |  |  |  |  | | | | | | | X | | | | | | X | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 10 | 13 |  |  |  |  |  | | X | | | X | X | X | | X | | | | X | | X | | | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | X | | | | | | | | | X | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | |
| 11 | 6 |  |  |  |  |  | | X | | | | X | X | | | | | | X | | | | | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | | X | | | | | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | |
| 12 | 6 |  |  |  |  |  | | X | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | | | | | X | X | X | | X | X | | | X | X | X | X | | | | X | | | X | X | X | X | X | X |
提供机构:
CyberHarem原始信息汇总
数据集概述
数据集名称
Dataset of matsubara_kanon/松原花音/마츠바라카논 (BanG Dream!)
数据集描述
该数据集包含500张图片及其标签,主题为matsubara_kanon/松原花音/마츠바라카논(BanG Dream!)。图片主要标签包括blue_hair, purple_eyes, bangs, long_hair, one_side_up, ribbon, hair_ribbon, hair_ornament, light_blue_hair。
数据集来源
图片从多个网站(如danbooru, pixiv, zerochan等)爬取,爬虫系统由DeepGHS Team提供支持。
数据集包列表
| 名称 | 图片数量 | 大小 | 下载链接 | 类型 | 描述 |
|---|---|---|---|---|---|
| raw | 500 | 683.74 MiB | Download | Waifuc-Raw | 包含元信息的原始数据(最小边对齐到1400像素,如果更大)。 |
| 800 | 500 | 399.39 MiB | Download | IMG+TXT | 短边不超过800像素的数据集。 |
| stage3-p480-800 | 1237 | 872.69 MiB | Download | IMG+TXT | 3阶段裁剪数据集,区域不小于480x480像素。 |
| 1200 | 500 | 610.99 MiB | Download | IMG+TXT | 短边不超过1200像素的数据集。 |
| stage3-p480-1200 | 1237 | 1.22 GiB | Download | IMG+TXT | 3阶段裁剪数据集,区域不小于480x480像素。 |
标签聚类结果
以下是部分标签聚类结果的示例:
原始文本版本
| # | 样本数量 | 图片示例 | 标签 |
|---|---|---|---|
| 0 | 15 | ![]() |
epaulettes, 1girl, band_uniform, shako_cap, solo, looking_at_viewer, open_mouth, green_ribbon, holding, sleeveless, wrist_cuffs, drumsticks, blush, white_skirt, :d, instrument |
| 1 | 44 | ![]() |
1girl, looking_at_viewer, solo, blush, bare_shoulders, short_sleeves, smile, hair_flower, frilled_sleeves, simple_background, white_background, blue_shirt, dress, upper_body, collarbone, green_ribbon, open_mouth, frilled_shirt, off_shoulder |
| 2 | 13 | ![]() |
1girl, blush, looking_at_viewer, solo, bow, short_sleeves, striped_ribbon, twintails, alternate_hairstyle, wrist_cuffs, earrings, red_ribbon, holding_gift, valentine, frilled_dress, center_frills, frilled_sleeves, gift_box, heart-shaped_box, heart_hair_ornament, white_background, :d, blue_ribbon, open_mouth, thighhighs |
| 3 | 11 | ![]() |
1girl, solo, earrings, looking_at_viewer, double_bun, blush, open_mouth, :d, dress, braid, jellyfish, long_sleeves, see-through_sleeves, shirt, hand_on_own_chest, upper_body |
| 4 | 7 | ![]() |
1girl, bow, detached_sleeves, looking_at_viewer, smile, solo, tiara, wrist_cuffs, blush, double_bun, frills, underwater, air_bubble, earrings, bun_cover, jellyfish, midriff, sidelocks, skirt, aqua_hair, bloomers, cleavage, gem, medium_breasts, open_mouth, pink_eyes, pink_ribbon |
| 5 | 35 | ![]() |
hanasakigawa_school_uniform, long_sleeves, 1girl, sailor_dress, blush, solo, brown_dress, red_ribbon, looking_at_viewer, neck_ribbon, smile, green_ribbon, open_mouth, double-breasted, upper_body, white_background, flower, simple_background, white_sailor_collar |
| 6 | 7 | ![]() |
1girl, bow, hair_flower, hairband, solo, white_flower, alternate_hairstyle, blush, floral_print, hair_over_shoulder, long_sleeves, ribbon_braid, single_braid, strawberry_hair_ornament, looking_at_viewer, print_dress, striped_ribbon, upper_body, white_gloves, :d, blue_dress, flower_earrings, open_mouth, pink_ribbon, upper_teeth_only, white_background, frilled_dress, frilled_shirt_collar, frilled_sleeves, red_ribbon |
| 7 | 7 | ![]() |
1girl, blush, hair_flower, long_sleeves, looking_at_viewer, solo, wide_sleeves, green_kimono, white_background, frilled_sleeves, maid_apron, white_flower, cosplay, frilled_apron, open_mouth, pom_pom_(clothes), wa_maid, white_apron, :d, green_bow, green_ribbon, hair_between_eyes, simple_background, striped_bow, upper_body |
| 8 | 6 | ![]() |
1girl, frills, looking_at_viewer, solo, blush, hat_bow, hat_ribbon, rabbit_ears, top_hat, chain, open_mouth, skirt, sleeveless, upper_body, :d, black_headwear, black_ribbon, blue_bow, heart_earrings, neck_ribbon, striped, watch, white_background, wrist_cuffs |
| 9 | 10 | ![]() |
2girls, blush, open_mouth, yuri, white_background, closed_eyes, blonde_hair |
| 10 | 13 | ![]() |
1girl, day, outdoors, solo, cloud, looking_at_viewer, bikini, blue_sky, blush, collarbone, open_mouth, cleavage, navel, :d, frills, medium_breasts, ocean, ball, holding, jewelry, standing |
| 11 | 6 | ![]() |
1girl, blush, hetero, solo_focus, 1boy, collarbone, large_breasts, looking_at_viewer, nipples, open_mouth, penis, clothing_aside, cum_in_pussy, sex, spread_legs, sweat, vaginal, bar_censor, bikini, girl_on_top, heavy_breathing, medium_breasts, mosaic_censoring, navel, sidelocks, straddling, tears |
| 12 | 6 | ![]() |
1girl, hetero, nipples, open_mouth, penis, sex, solo_focus, heart-shaped_pupils, medium_breasts, navel, pussy, spread_legs, 1boy, completely_nude, cum, mosaic_censoring, nose_blush, sweat, testicles, vaginal |
搜集汇总
数据集介绍

构建方式
该数据集聚焦于《BanG Dream!》中的角色松原花音,整合了来自Danbooru、Pixiv、Zerochan等多个图像平台的500张图片及其对应标签。数据采集由DeepGHS团队开发的自动化爬取系统完成,确保来源的多样性与覆盖广度。在构建过程中,针对角色的核心标签(如蓝发、紫瞳、刘海、长发等)进行了精简处理,以突出角色辨识特征。数据以多种分辨率版本提供,包括原始版(长边对齐至1400像素)、800与1200像素版本,以及三阶段裁剪版(面积不低于480×480像素),满足不同训练需求。
特点
该数据集的一大特色在于其精细的标签聚类结果,通过将图像按标签相似度分组,可挖掘出角色在不同场景下的服饰与装扮模式,例如校服、泳装、和服等风格簇。每个聚类均附带样本图像与标签列表,为风格迁移或细粒度图像生成提供结构化参考。此外,数据集内含的原始元数据(如文件名与标签)支持灵活的定制加载,而裁剪版本则通过多阶段处理优化了图像构图,提升模型训练效率。整体而言,数据集兼顾了数据规模(500张)与多样性,适合小样本学习场景。
使用方法
数据集支持通过waifuc库进行高效加载,用户可下载原始压缩包后,利用LocalSource接口逐项读取图像及其元数据。具体流程包括从HuggingFace Hub下载dataset-raw.zip文件,解压至本地目录,再通过waifuc的迭代器访问图像、文件名与标签信息。此外,用户亦可直接选用预处理的IMG+TXT格式版本(如800或1200像素),无需额外标注即可用于文本到图像模型的训练。聚类数据以表格形式呈现,便于手动分析或作为数据增强的参考依据。
背景与挑战
背景概述
在文本到图像生成领域,高质量、结构化的角色数据集是驱动模型精准生成特定动漫角色的关键。CyberHarem/matsubara_kanon_bangdream 数据集由 DeepGHS 团队于近期构建,聚焦于《BanG Dream!》中的角色松原花音。该数据集的核心研究问题在于如何系统性地收集、标注和呈现动漫角色的多姿态、多服装图像,以支持生成模型的细粒度学习。通过从 Danbooru、Pixiv 等多个平台自动爬取,并配备基于 Waifuc 的标准化处理流程,该数据集提供了 500 张原始图像及其标签,并衍生出多种分辨率版本与裁剪版本,为角色生成任务提供了可复现的基准数据源。其在 Hugging Face 上的发布,促进了动漫角色生成研究的开放协作,对相关领域的社区贡献显著。
当前挑战
该数据集面临的挑战主要体现在两个层面。在领域问题层面,尽管数据集聚焦于单一角色,但动漫角色生成仍面临身份一致性难题:如何在多服装、多场景下保持松原花音的独特视觉特征(如蓝发紫瞳、单侧发饰)不变形,是文本到图像模型的核心挑战。在构建过程中,挑战则源于数据来源的异构性与质量控制:从多个网站自动爬取的图像在分辨率、构图、标签一致性上差异显著,需通过裁剪策略(如 stage3-p480-800)与聚类分析(如按服装分组)来缓解噪声。此外,数据集中包含少量成人内容(如标签中的“hetero”“nipples”),如何在不违反平台政策的前提下进行合理标注与过滤,亦构成伦理与工程上的双重挑战。
常用场景
经典使用场景
该数据集专为文本到图像生成任务而设计,聚焦于BanG Dream!中角色松原花音的视觉呈现。其核心应用在于通过提供的500张高质量图像及精细化标签(如蓝发、紫瞳、刘海等特征),驱动生成式模型学习该角色的外貌风格与服饰细节。研究者常利用此数据集进行角色一致性生成实验,例如在Stable Diffusion等模型上微调,以实现对特定二次元角色的精准还原。
解决学术问题
在学术领域,该数据集有效解决了二次元角色数据稀缺且标注不统一的难题。它通过标准化标签体系(如服饰、发型、姿态等维度)和多种分辨率版本(如800px、1200px),为图像生成中的细粒度控制研究提供了基准。其意义在于推动了角色专属生成模型的发展,使学者能够系统性地探索标签质量与生成保真度之间的关联,并验证数据增强策略(如三阶段裁剪)对模型泛化能力的影响。
衍生相关工作
该数据集衍生出多项经典工作,包括基于Waifuc框架的自动化图像采集与标注流水线,以及针对角色多视角一致性的3D感知生成方法。相关研究还探索了标签聚类在风格解耦中的应用,例如通过分析不同簇(如泳装簇、和服簇)的标签分布,实现角色属性与场景的分离生成。这些工作共同推动了二次元图像生成领域的标准化与工具化进程。
以上内容由遇见数据集搜集并总结生成
















