LS-HDIB

LS-HDIB是一个大规模的手写文档图像二值化数据集，由印度理工学院甘地分校CVIG实验室创建。该数据集包含超过一百万张手写文档图像，覆盖了多种真实世界场景。数据集的创建结合了自适应阈值和无缝克隆技术，确保了高质量的分割手写内容。LS-HDIB数据集特别适用于深度学习模型训练，以提高在复杂背景和多种手写风格下的文档图像二值化性能。

LS-HDIB is a large-scale handwritten document image binarization dataset created by the CVIG Lab at the Indian Institute of Technology Gandhinagar. This dataset contains over one million handwritten document images covering diverse real-world scenarios. It was developed using a combination of adaptive thresholding and seamless cloning techniques to ensure high-quality segmentation of handwritten content. The LS-HDIB dataset is particularly suitable for training deep learning models to improve their document image binarization performance under complex backgrounds and various handwriting styles.