基于YOLOv8的手写甲骨文检测算法优化  

Handwritten Oracle Bone Inscriptions Detection Algorithm Based on YOLOv8

作　　者：陶筱娇 李健[1] 王帅[2] 杨钧 王永归 TAO Xiao-jiao;LI Jian;WANG Shuai;YANG Jun;WANG Yong-gui(School of Electronic Information and Artificial Intelligence,Shaanxi University of Science and Technology,Xi’an 710021,China;School of History and Civilization,Shaanxi Normal University,Xi’an 710119,China)

机构地区：[1]陕西科技大学电子信息与人工智能学院,西安710021 [2]陕西师范大学历史文化学院,西安710119

出　　处：《印刷与数字媒体技术研究》2025年第2期194-200,共7页Printing and Digital Media Technology Study

基　　金：2018年度国家社科基金西部项目(No.18XKG003)。

摘　　要：为解决手写甲骨文人工检测效率低、识别精度不足的问题,本研究将深度学习算法应用于手写甲骨文的自动检测中。首先,建立了手写甲骨文数据集并对其进行标注;然后,通过手动扩充数据集和二次标注,提高了数据集的可信度;最后,在YOLOv8算法基础上,使用滑窗裁剪和识别辅助技术对检测算法进行优化,提高甲骨文检测的准确性。实验结果表明,本研究提出的优化算法提高了模型检测手写甲骨文的精度,鲁棒性较好,具有一定的研究价值和应用前景。To solve the problems of low manual detection efficiency and insufficient identification accuracy,the deep learning algorithm was applied to automatical detect handwritten oracle bone inscriptions in this study.Firstly,a dataset of handwritten oracle bone inscriptions was established and annotated.Then,the dataset was manually expanded and re-annotated to improve its credibility.Finally,based on the YOLOv8 algorithm,the sliding window cropping and recognition-assisted techniques were used to optimize the detection algorithm and improve the accuracy of oracle bone inscription detection.The experimental results demonstrated that the optimization algorithm proposed in this study improves the precision of model detection for handwritten oracle bone inscriptions,exhibiting good robustness,which has certain research value and application prospects.

关 键 词：甲骨文检测 深度学习 目标检测 YOLOv8 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]