TRR360D

TRR360D是由上海电力大学电子与信息工程学院创建的旋转表格检测数据集，旨在解决传统表格检测数据集稀缺和高标注成本的问题。该数据集基于ICDAR2019MTD数据集，参考DOTA数据集的标注格式，包含600张训练图像和977个标注实例，以及240张测试图像和499个标注实例。数据集通过约束起始点和标注方向来创建，适用于研究旋转表格检测算法，推动表格检测技术的发展。

TRR360D is a rotated table detection dataset created by the School of Electronics and Information Engineering, Shanghai University of Electric Power. It aims to address the issues of scarcity and high annotation cost in traditional table detection datasets. Built upon the ICDAR2019MTD dataset and referencing the annotation format of the DOTA dataset, this dataset contains 600 training images with 977 annotation instances, as well as 240 test images with 499 annotation instances. Constructed by constraining the starting points and annotation directions, the dataset is suitable for research on rotated table detection algorithms and facilitates the advancement of table detection technologies.