遇见数据集

Ancient-3

收藏
arXiv2025-09-30 收录
官方服务:

资源简介:

该数据集名为Ancient-3,包含了1,186个类别,共39,009个手写甲骨文样本,涵盖了早期的书写系统。此外,该数据集面临着数据不平衡和类别内变异的挑战。规模上,该数据集包含了39,009张图像,其任务是进行甲骨文字符识别。

This dataset, designated as Ancient-3, contains 39,009 handwritten oracle bone script samples across 1,186 categories, covering early writing systems. Furthermore, this dataset confronts two major challenges: data imbalance and intra-class variation. With a total of 39,009 images, the core task of this dataset is oracle bone character recognition.

二维码
社区交流群
二维码
科研交流群
商业服务