DCM 漫画图像数据集

DCM 数据集由 772 张带标注的图片组成，图片来自 27 本黄金时代的漫画书。数据搜集自 Digital Comics Museum 。该数据集从每个可用的出版商中选择一个专辑，以获得尽可能多的不同风格。该数据集为所有的面板和所有的人物（身体+脸部）标注了 ground truth 边界框，这些人物或小或大、或像人或像动物。

The DCM dataset comprises 772 annotated images derived from 27 Golden Age comic books, with all data collected from the Digital Comics Museum. To obtain the broadest possible range of artistic styles, the dataset selects one comic issue from each available publisher. Ground-truth bounding boxes are annotated for every comic panel and all characters (encompassing both their bodies and faces), including figures of varying sizes that represent both humans and animals.

搜集汇总

背景与挑战

背景概述

DCM 漫画图像数据集包含772张来自27本黄金时代漫画书的标注图片，数据源自Digital Comics Museum，涵盖了多种绘画风格。该数据集为所有面板和人物（包括身体和脸部）提供了ground truth边界框标注，适用于人脸识别任务。