KOHTD
收藏资源简介:
KOHTD(Kazakh Offline Handwritten Text Dataset)是由阿尔-法拉比哈萨克国立大学和Satbayev大学合作创建的一个大型手写文本数据集,专门用于哈萨克语的手写文本识别研究。该数据集包含3000份手写考试卷,总计超过140,335张分割图像和约922,010个符号。数据集的创建过程涉及扫描、预处理、分割和标注等多个步骤,确保了数据的质量和可用性。KOHTD数据集主要用于深度学习和机器学习领域的研究,特别是在手写文本识别任务中,为算法性能的评估和比较提供了标准化的数据资源。
KOHTD (Kazakh Offline Handwritten Text Dataset) is a large-scale handwritten text dataset jointly created by Al-Farabi Kazakh National University and Satbayev University, specifically dedicated to handwritten text recognition research for the Kazakh language. This dataset contains 3000 handwritten exam papers, totaling over 140,335 segmented images and approximately 922,010 symbols. The creation process of the KOHTD dataset involves multiple steps such as scanning, preprocessing, segmentation and annotation, which ensures the data quality and usability. The KOHTD dataset is mainly used for research in the fields of deep learning and machine learning, especially in handwritten text recognition tasks, providing standardized data resources for the evaluation and comparison of algorithm performance.




