hal-utokyo/MangaOCR
收藏官方服务:
资源简介:
这是一个名为MangaOCR的数据集,由Manga109数据集和漫画拟声词数据集的注释合并构建而成。它包含了大约209K个叙事文本实例,涵盖了多种视觉风格和布局。
This is the MangaOCR dataset. We constructed this dataset by consolidating annotations from the Manga109 dataset and the manga onomatopoeia dataset. It contains approximately 209K narrative text instances, spanning a wide variety of visual styles and layouts.
提供机构:
hal-utokyo


