five

Old Babylonian cuneiform signs from Akkadian documentary tablets

收藏
Zenodo2026-02-09 更新2026-05-26 收录
下载链接:
https://zenodo.org/doi/10.5281/zenodo.18335468
下载链接
链接失效反馈
官方服务:
资源简介:
The following dataset consists of a collection of sign annotations gathered from Old Babylonian cuneiform documentary tablets. The tablets have been acquired with White Light Portable Light Domes that created 2D+ data files. From these files, various renders have been made: photorealistic with even illumination and various lighting angles, and non-photorealistic based on depth information. In this dataset there are data of one photorealistic visualisation with lighting angled from above (ColorA) and one non-photorealistic sketch visualisation (SketchB).  Three expert annotators have annotated each visible cuneiform sign on 303 cuneiform tablets. The annotations are originally made with polygonal borders and from these square annotations are made based on the longest side. Annotation files are JPGs and named after the tablet (e.g., HS2344), the surface (e.g., front or bottom) and lastly they are given a random ID. The signs are grouped in folders identifying their readings (e.g., asz or na). The file names are [reading]/[tablet name]_[face]_[id]_[visualization].jpg. The visualization part can be used to filter the data for only a single visualization. See e.g. https://github.com/VerwimpEli/signs-of-the-past for code to create dataset ready for machine learning experiments to use this data. The file structure is as follows: sign_crops |--- 1(ban2) |    |---- CUNES51-01-89_back_983b4c00af2f_ColorA.jpg |    |---- CUNES51-01-89_back_983b4c00af2f_SketchB.jpg |    |---- CUNES51-01-89_back_c1b3cb520f82_ColorA.jpg |    |---- CUNES51-01-89_back_c1b3cb520f82_SketchB.jpg |    ... |--- 1(gur) |    |---- HS_2186_back_fba1d15da2ea_ColorA.jpg |    |---- HS_2186_back_fba1d15da2ea_SketchB.jpg |    ... ... Accompanying the dataset is a index of the different sign categories attested in the corpus with their variants and number of attestations. This file can be used to combine different readings into a single visual category. The annotations are all made in the Cuneur annotation platform (introduced in Homburg et al. 2022. "Annotated 3D-Models of Cuneiform Tablets," in Journal of Open Archaeology Data vol. 10(92)) The folder with the annotations is called: sign_cropsIndex file is called: sign_map.csv (Note: this file uses ';' as a delimiter, as the default ',' is used in some of the ASCII readings of cuneiform signs)
提供机构:
Zenodo
创建时间:
2026-02-09
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作