翻译：在甲骨学专家的指导下到目前为止数据集整理团队包括2位老师和10名同学对20册甲骨文献图像中的2738张进行了字符检测标注即数据集总共包括 2738张图像在这些文献图像中用矩形框标注了19595个甲骨文字符平均每页甲骨文献图像包括 7个左右的字符示例。本次标注将甲骨文献中出现的古文字标为正样本标签为1。从标注字符的角度看甲骨文献中的字符对象基本是小目标且呈现显著尺度多样性因此甲骨字符检测归属于

Under the guidance of experts in oracle bone script, the data set organizing team, consisting of two teachers and ten students, has so far annotated 2738 images of oracle bone script documents, including a total of 19595 characters marked with rectangular boxes. On average, each page of oracle bone script documents includes about 7 character examples. In this annotation, ancient characters appearing in oracle bone script documents were marked as positive samples with a label of 1. From the perspective of character annotation, the character objects in oracle bone script documents are mostly small targets and exhibit significant scale diversity. Therefore, oracle character detection belongs to small visual target detection and requires consideration of multi-scale character detection, which should be a key factor to consider in the design process of detection methods