Dead Sea Scrolls data collection (images, labels, prediction plots) for dating ancient manuscripts using radiocarbon and AI-based writing style analysis

The dataset is associated with the following article:Title: Dating ancient manuscripts using radiocarbon and AI-based writing style analysisAuthors: Mladen Popović, Maruf A. Dhali, Lambert Schomaker, Johannes van der Plicht, Kaare Lund Rasmussen, Jacopo La Nasa, Ilaria Degano, Maria Perla Colombini, and Eibert Tigchelaar(Under review) This data set is collected for the ERC project:The Hands that Wrote the Bible: Digital Palaeography and Scribal Culture of the Dead Sea ScrollsPI: Mladen PopovićGrant agreement ID: 640497Project website: https://cordis.europa.eu/project/id/640497 Copyright (c) University of Groningen, 2024. All rights reserved.Disclaimer and copyright notice for all data contained on the *.tar.gz files: 1) permission is hereby granted to use the data for research purposes. It is not allowed to distribute this data for commercial purposes. 2) provider gives no express or implied warranty of any kind, and any implied warranties of merchantability and fitness for purpose are disclaimed. 3) provider shall not be liable for any direct, indirect, special, incidental, or consequential damages arising out of any use of this data. 4) the user should refer to the first public article mentioned above on this data set. 5) the recipient should refrain from proliferating the data set to third parties external to his/her local research group. Please refer interested researchers to this site to obtain their own copy. Organization of the data:(Update on 19 April 2024: OxCal data for accepted 2-sigma ranges are updated with the incusion and exclusion of minor peaks. New prediction plots are added after the model is trained with accepted 2-sigma ranges, including minor peaks. The old plots are also kept. <OLD Updates below; disregard>updated on 07-Feb-2024: OxCal data for selected ranges added in a new directory in addition to previously available original OxCal data. Enoch's prediction plots and test images are reorganized for easy access to the users.<OLD Updates above; disregard>Please use the files from this version and disregard the previous two versions: 10.5281/zenodo.10629480 and 10.5281/zenodo.8168210) There are four *.tar.gz files: C14-Oxcal-data-updated.tar.gz contains one directory with radiocarbon data (OxCal [1] raw data) for all 30 manuscripts. Three additional directories contain name-corrected files for original OxCal data, files with accepted ranges, and files with accepted ranges including minor peaks. Please refer to the original article for details about OxCal data and the manuscripts. 25 out of 30 raw OxCal data are used (accepted ranges only) as the training labels during the training of Enoch, the date prediction model. train-images-c14.tar.gz contains the clean and preprocessed (binarized, aligned, and arrangement corrected) training images for the 25 radiocarbon-dated training manuscripts (including 4Q52; 64 images in total). test-images-all.tar.gz contains the clean and preprocessed test images for 135 previously undated manuscripts. The images are organized in three different directories: the first one with all 359 images for the 135 manuscripts, the second one with the selected 135 images, and the final one with 25 images to illustrate the poor quality of images. Enoch-prediction-new-with-minor-peaks.tar.gz contains the new date prediction plots for each of the 135 test images, where Enoch was trained with the inclusion of minor peaks for the 2-sigma accepted ranges and with a data balancing threshold of 0.05. These plots are used by expert palaeographers' evaluation of Enoch's style-based date predictions of 135 previously undated manuscripts. Enoch-predictions.tar.gz contains the date prediction plots for each of the 135 test images. There are two directories inside the *.tar.gz file:- prediction-plots-for-selected-135: Prediction plots with data balancing threshold of 0.05. - extra-plots: contains four additional directories: - Enoch-predictions-c14wo4Q52-balanced05: Prediction plots with data balancing threshold of 0.05. - Enoch-predictions-c14wo4Q52-balanced10: Prediction plots with data balancing threshold of 0.1. - Enoch-predictions-c14wo4Q52-unbalanced: Unbalanced raw predictions. - Enoch-predictions-c14wo4Q52-combined: Combined plots with all three prediction plots (unbalanced, 0.05, 0.1).Please refer to the original article for more details. The updated code to run the plot is available here: https://doi.org/10.5281/zenodo.10998860 If you have any questions, please get in touch with us:Mladen Popović <m.popovic(at)rug.nl>Maruf A. Dhali <m.a.dhali(at)rug.nl>Lambert Schomaker <l.r.b.schomaker(at)rug.nl> References:1. Bronk Ramsey, C. (2001). Development of the radiocarbon calibration program. Radiocarbon, 43(2A), 355-363.本数据集关联以下论文： 标题：基于放射性碳定年与AI书写风格分析的古手稿断代 作者：Mladen Popović、Maruf A. Dhali、Lambert Schomaker、Johannes van der Plicht、Kaare Lund Rasmussen、Jacopo La Nasa、Ilaria Degano、Maria Perla Colombini 与 Eibert Tigchelaar（处于审稿阶段） 本数据集为欧洲研究理事会（European Research Council, ERC）项目「书写圣经之手：死海古卷的数字古文字学与抄写文化」所采集，项目负责人为Mladen Popović，资助协议编号：640497，项目官网：https://cordis.europa.eu/project/id/640497 版权所有 © 格罗宁根大学，2024。保留所有权利。以下为所有*.tar.gz文件包含数据的免责声明与版权须知： 1. 特此授予研究用途使用本数据的权限。不得出于商业目的分发本数据。 2. 提供者不提供任何明示或默示担保，特此排除任何针对适销性及特定用途适用性的默示担保。 3. 对于因使用本数据所引发的任何直接、间接、特殊、偶然或后果性损害，提供者概不承担责任。 4. 用户应引用上文提及的本数据集首篇公开论文。 5. 接收方不得将本数据集扩散至其本地研究团队以外的第三方。如有研究人员希望获取本数据集，请指引其访问本网站自行下载。 【数据更新说明（2024年4月19日更新）：已针对经确认的2σ范围的OxCal数据，通过纳入与排除次要峰进行了更新。在使用经确认的2σ范围（含次要峰）训练模型后，新增了预测绘图，同时保留了旧版绘图。<以下为旧版更新，可忽略>2024年2月7日更新：在现有原始OxCal数据目录外，新增包含选定范围OxCal数据的新目录；对Enoch的预测绘图与测试图像进行了重组，以方便用户访问。<以上为旧版更新，可忽略>请使用本版本的文件，忽略此前两个版本：10.5281/zenodo.10629480 与 10.5281/zenodo.8168210】 本数据集包含4个*.tar.gz文件： 1. C14-Oxcal-data-updated.tar.gz：内含1个存放30份手稿放射性碳数据（OxCal[1]原始数据）的目录，另有3个附加目录，分别存放经文件名校正后的原始OxCal数据文件、仅含经确认范围的数据文件，以及包含次要峰的经确认范围数据文件。有关OxCal数据与手稿的详细信息，请参阅原论文。训练Enoch（断代预测模型）时，共使用30份原始OxCal数据中的25份（仅采用经确认的范围）作为训练标签。 2. train-images-c14.tar.gz：内含针对25份经放射性碳定年的训练手稿（含4Q52）的经预处理（二值化、对齐与排版校正）的干净训练图像，总计64张。 3. test-images-all.tar.gz：内含针对135份此前未断代手稿的经预处理干净测试图像。图像分为3个目录：第一个目录包含135份手稿的全部359张图像；第二个目录包含精选的135张图像；第三个目录包含25张用于展示图像质量不佳的示例图像。 4. Enoch-prediction-new-with-minor-peaks.tar.gz：内含针对135张测试图像的新版断代预测绘图。该模型训练时纳入了次要峰的2σ经确认范围，并采用0.05的数据平衡阈值。这些绘图将用于古文字学专家对Enoch针对135份此前未断代手稿的风格导向断代预测结果的评估。 5. Enoch-predictions.tar.gz：内含针对135张测试图像的断代预测绘图。该压缩包内含2个目录： - prediction-plots-for-selected-135：采用0.05的数据平衡阈值生成的预测绘图。 - extra-plots：包含4个附加目录： - Enoch-predictions-c14wo4Q52-balanced05：采用0.05的数据平衡阈值生成的预测绘图。 - Enoch-predictions-c14wo4Q52-balanced10：采用0.1的数据平衡阈值生成的预测绘图。 - Enoch-predictions-c14wo4Q52-unbalanced：未平衡的原始预测结果。 - Enoch-predictions-c14wo4Q52-combined：包含全部三种预测绘图（未平衡、0.05阈值、0.1阈值）的组合绘图。 详细信息请参阅原论文。 运行绘图的更新代码可在此获取：https://doi.org/10.5281/zenodo.10998860 如有任何疑问，请联系： Mladen Popović <m.popovic(at)rug.nl> Maruf A. Dhali <m.a.dhali(at)rug.nl> Lambert Schomaker <l.r.b.schomaker(at)rug.nl> 参考文献： 1. Bronk Ramsey, C. (2001). 放射性碳定年校准程序的开发. 放射性碳, 43(2A), 355-363.