摘要：【中文】本发明涉及一种数据DNA可视化关系分析系统及方法，属于大数据应用技术领域，包括：系统管理模块，用于完成系统的基础配置和维护；标准化管理模块，用于提供定义所管理的元数据相关的标准，提供可视化的稽核报告；元数据管理模块，用于实现元数据的管理；流程管理模块，用于实现上线相关功能；调度管理模块，用于管理调度环境、配置调度任务，实现采集数据、加工处理及数据共享；运维监控模块，用于对系统进行监控管理；数据质量模块，用于对所管理数据的质量进行评估管理；数据目录模块，对所管理的对象按用户配置的目录进行分门别类地显示和管理，使数据更清晰；生命周期管理模块，用于对系统所管理的对象的生命周期进行管理。 【EN】The invention relates to a system and a method for analyzing a data DNA visualization relationship, which belong to the technical field of big data application and comprise the following steps: the system management module is used for completing basic configuration and maintenance of the system; the standardized management module is used for providing standards related to metadata which are defined and managed and providing visual audit reports; the metadata management module is used for realizing the management of metadata; the flow management module is used for realizing the related online function; the scheduling management module is used for managing a scheduling environment, configuring scheduling tasks and realizing data acquisition, processing and data sharing; the operation and maintenance monitoring module is used for monitoring and managing the system; the data quality module is used for evaluating and managing the quality of the managed data; the data directory module displays and manages the managed objects in a classified manner according to the directory configured by the user, so that the data are clearer; and the life cycle management module is used for managing the life cycle of the object managed by the system.

摘要：【中文】本发明涉及一种基于多种神经网络混合的出版PDF版面分析识别方法，属于图像识别、PDF版面分析技术领域，采用多任务训练模式，先对版面进行识别，分割并标注PDF版面，包括段落、标题、插图，并定位文本行，再对文本进行识别。本方法在版面识别上，通过多任务训练模式，同时完成了本文行、结构识别标注，全过程无需人工参与，有效保留了PDF文本结构信息。根据版面分析得到的带PDF文本结构信息的数据，构建出版数据常用中文字典，并针对性训练文本识别模型，从而使模型在PDF印刷体文本识别任务中识别精度得以大幅提升。识别后的文本同样具有结构信息，还原原始PDF布局结构，也便于后续二次编辑，制作电子书，挖掘图书内容知识。 【EN】The invention relates to a publishing PDF layout analysis and identification method based on mixing of multiple neural networks, which belongs to the technical field of image identification and PDF layout analysis. In the method, the text line and structure identification and marking are simultaneously completed through a multi-task training mode on the aspect of layout identification, manual participation is not needed in the whole process, and the PDF text structure information is effectively reserved. According to the data with the PDF text structure information obtained through layout analysis, a publishing data common Chinese dictionary is constructed, and a text recognition model is trained in a targeted mode, so that the recognition accuracy of the model in a PDF print text recognition task is greatly improved. The recognized text also has structural information, an original PDF layout structure is restored, subsequent secondary editing is facilitated, an electronic book is manufactured, and book content knowledge is mined.