Word-Zone Roi Ids+Txt For The Monk Historical Manuscript Collection / June 2018
收藏资源简介：
Checkpointed state of the labeling situation in the Monk e-Science service for handwritten document indexing as of June 2018. The reason for the upload is the migration from the IBM gpfs system of the Target project (EU/SNN) to a new Lustre file system of the CIT of the university of Groningen. This valuable indexing information contains 'the human' labels for regions of interest on page scans. However, the original images are needed to make sense of this. In case of a calamity, they very often can be reconstructed because the orginal scan file names by the archives and institutions are maintained. The amount of image data would be too much for uploading to Zenodo, too. Example data portion: Monk internal ID with embeded x,y,w,h of ROIs,<br> followed by <txt>label....</txt> <br> Word zones are of the type RECOG or HUMAN. <sub>navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=0285-y=0359-w=1216-h=<br> 0094-ybas=0433-nink=174-ns=2-segm=COCOS8cocos <txt>0,</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=0285-y=0430-w=0020-h=<br> 0023-ybas=0433-nink=114-ns=1-segm=COCOS8cocos <txt>x</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0243-w=0339-h=<br> 0215-ybas=0433-nink=7762-ns=8-segm=COCOS8cocos <txt>Dadap</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0243-w=0384-h=<br> 0215-ybas=0433-nink=7690-ns=3-segm=COCOS8cocos <txt>Dadap</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0243-w=0396-h=<br> 0215-ybas=0433-nink=8614-ns=9-segm=COCOS8cocos <txt>Mandor's</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0244-w=0237-h=<br> 0214-ybas=0433-nink=7758-ns=7-segm=COCOS8cocos <txt>@Tropinatus</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0245-w=0228-h=<br> 0213-ybas=0433-nink=7708-ns=6-segm=COCOS8cocos <txt>Pati</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0246-w=0223-h=<br> 0212-ybas=0433-nink=7705-ns=5-segm=COCOS8cocos <txt>@Tropinatus</txt></sub>
本数据集为2018年6月时，Monk电子科学服务中用于手写文档索引的标注任务检查点状态。本次上传的缘由是Target项目（EU/SNN）将其存储系统从IBM gpfs迁移至格罗宁根大学CIT的全新Lustre文件系统。 该极具价值的索引信息包含了页面扫描图像中感兴趣区域（ROIs，Regions of Interest）的人工标注（human labels）结果。但若要准确理解该标注数据，则需结合原始扫描图像方可实现。即便遭遇数据灾变，由于档案机构与相关单位均保留了原始扫描文件的命名规则，标注数据通常可被重建。此外，原始图像数据体量过大，同样不适用于上传至Zenodo平台。 示例数据片段如下： 包含嵌入感兴趣区域坐标x、y、w、h的Monk内部标识符，其后跟随`<txt>标注内容</txt>`格式的标签信息。 字词区域分为RECOG与HUMAN两种类型。 <sub>navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=0285-y=0359-w=1216-h=<br> 0094-ybas=0433-nink=174-ns=2-segm=COCOS8cocos <txt>0,</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=0285-y=0430-w=0020-h=<br> 0023-ybas=0433-nink=114-ns=1-segm=COCOS8cocos <txt>x</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0243-w=0339-h=<br> 0215-ybas=0433-nink=7762-ns=8-segm=COCOS8cocos <txt>Dadap</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0243-w=0384-h=<br> 0215-ybas=0433-nink=7690-ns=3-segm=COCOS8cocos <txt>Dadap</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0243-w=0396-h=<br> 0215-ybas=0433-nink=8614-ns=9-segm=COCOS8cocos <txt>Mandor's</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0244-w=0237-h=<br> 0214-ybas=0433-nink=7758-ns=7-segm=COCOS8cocos <txt>@Tropinatus</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0245-w=0228-h=<br> 0213-ybas=0433-nink=7708-ns=6-segm=COCOS8cocos <txt>Pati</txt><br> navis-NNM001001033_0001-line-001-y1=44-y2=552-zone-RECOG-x=1487-y=0246-w=0223-h=<br> 0212-ybas=0433-nink=7705-ns=5-segm=COCOS8cocos <txt>@Tropinatus</txt></sub>
