KVP10k
收藏arXiv2024-05-01 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2405.00505v1
下载链接
链接失效反馈官方服务：
资源简介：
KVP10k是由IBM研究院以色列分部和苏黎世分部联合创建的一个全面的数据集，专注于商业文档中的关键值对提取。该数据集包含10707个丰富注释的图像，旨在解决非预定义关键值对提取的挑战。数据集通过广泛的网络爬虫和公共文件收集，确保了数据来源的多样性和真实性。KVP10k的应用领域广泛，包括但不限于自动化数据录入、快速信息检索和决策支持，旨在提高企业运营效率和竞争力。
KVP10k is a comprehensive dataset jointly developed by the Israel and Zurich Research Labs of IBM, focusing on key-value pair extraction from business documents. It contains 10,707 richly annotated images, and is designed to address the challenges of extracting non-predefined key-value pairs. The dataset is collected via extensive web crawling and public document sources, ensuring the diversity and authenticity of its data sources. KVP10k has a wide range of application scenarios, including but not limited to automated data entry, rapid information retrieval and decision support, aiming to enhance enterprise operational efficiency and competitiveness.
提供机构：
IBM研究院以色列分部，海法大学校区，卡梅尔山，海法3498825，以色列 IBM研究院苏黎世分部，Säumerstrasse 4，8803 Rüschlikon，瑞士
创建时间：
2024-05-01
