Innovative Applications and Improvements of Deep Learning in the Detection and Recognition of Oracle-bone Character Images

WANG Zexin; TONG Hengjian

doi:10.11907/rjdk.251006

您当前的位置：

首页 >

文章列表页 >

Innovative Applications and Improvements of Deep Learning in the Detection and Recognition of Oracle-bone Character Images

Graphic and Image Processing | 更新时间：2026-04-30

- Innovative Applications and Improvements of Deep Learning in the Detection and Recognition of Oracle-bone Character Images
- Software Guide Vol. 25, Issue 4, Pages: 162-167(2026)
- 作者机构：
  
  中国地质大学（武汉）计算机学院，湖北武汉 430074
- 作者简介：
- 基金信息：
- DOI：10.11907/rjdk.251006
  CLC： TP391
- Received：03 January 2025，
  
  Published：15 April 2026
- 稿件说明：
移动端阅览
王泽鑫,童恒建.基于深度学习的甲骨文图像检测与识别创新应用及改进研究[J].软件导刊,2026,25(04):162-167.

WANG Zexin,TONG Hengjian.Innovative Applications and Improvements of Deep Learning in the Detection and Recognition of Oracle-bone Character Images[J].Software Guide,2026,25(04):162-167.
王泽鑫,童恒建.基于深度学习的甲骨文图像检测与识别创新应用及改进研究[J].软件导刊,2026,25(04):162-167. DOI： 10.11907/rjdk.251006.

WANG Zexin,TONG Hengjian.Innovative Applications and Improvements of Deep Learning in the Detection and Recognition of Oracle-bone Character Images[J].Software Guide,2026,25(04):162-167. DOI： 10.11907/rjdk.251006.

摘要

甲骨文图像样本量稀少、细节复杂，且分为临摹和拓片两种图像，其检测与识别难度较大。为此，提出一种基于深度学习的甲骨文图像检测与识别方案。具体来说，针对样本量稀少的问题，提出字体图像类的数据增强WDA算法和拓片化处理过程；针对拓片甲骨文字检测数据集的空白，构建Oracle-Detect数据集；针对传统模型识别率低、难以增量识别的问题，提出基于Stacking的集成式甲骨文字识别模型。实验结果表明，该方案对甲骨文的检测mAP达到94%，在临摹甲骨文字识别任务上的准确率达到83%，对拓片甲骨文字的识别准确率达到81%。

Abstract

The sample size of oracle bone script images is scarce， with complex details and divided into two types： sketches and rubbings， making their detection and recognition particularly challenging. To address this， a deep learning-based detection and recognition solution for oracle bone script images is proposed. Specifically， to tackle the issue of limited sample sizes， a font-image class data augmentation algorithm （WDA） and a rubbing processing procedure are introduced. To fill the gap in the dataset for detecting oracle bone script on rubbings， the Oracle-Detect dataset is constructed. To overcome the low recognition rates and difficulties in incremental recognition of traditional models， a Stacking ensemble-based oracle bone script recognition model is improved. Experimental results demonstrate that this solution achieves a detection mAP of 94% for oracle bone script， an accuracy rate of 83% in recognizing sketch-based oracle bone script， and an accuracy rate of 81% in identifying rubbing-based oracle bone script.

关键词

Keywords

references

XIE N H . Research on oracle bone inscriptions， a "unique learning" moving from Yin Ruins to the world——a review of the "International Academic Symposium Commemorating the 120th Anniversary of the Discovery of Oracle Bone Inscriptions" in Busan， South Korea ［J］. Guanzi Journal ， 2020 （ 3 ）： 125 - 128 . doi: 10.19321/j.cnki.gzxk.issn1002-3828.2020.03.16 http://dx.doi.org/10.19321/j.cnki.gzxk.issn1002-3828.2020.03.16

谢乃和 . 从殷墟走向世界的“绝学”甲骨文字研究——韩国釜山“纪念甲骨文发现120周年国际学术研讨会”述评［J］. 管子学刊， 2020 （ 3 ）： 125 - 128 . doi: 10.19321/j.cnki.gzxk.issn1002-3828.2020.03.16 http://dx.doi.org/10.19321/j.cnki.gzxk.issn1002-3828.2020.03.16

LIU G Y . Based on deep learning oracle-bone character detection and recognition ［J］. Yindu Journal ， 2020 ， 41 （ 3 ）： 54 - 59 . doi: 10.3969/j.issn.1001-0238.2020.03.010 http://dx.doi.org/10.3969/j.issn.1001-0238.2020.03.010

刘国英 . 基于深度学习的甲骨文字检测与识别［J］. 殷都学刊， 2020 ， 41 （ 3 ）： 54 - 59 . doi: 10.3969/j.issn.1001-0238.2020.03.010 http://dx.doi.org/10.3969/j.issn.1001-0238.2020.03.010

TU C N ， WANG G ， TIAN J ， et al . Research on classification algorithms of oracle-bone characters based on deep learning ［J］. Modern Computer ， 2021 ， 27 （ 26 ）： 67 - 72 .

涂淳宁，王贵，田吉，等 . 基于深度学习的甲骨文分类算法研究［J］. 现代计算机， 2021 ， 27 （ 26 ）： 67 - 72 .

LI F ， ZHOU X L . Graph Theory method for automatic recognition of oracle bone inscriptions ［J］. Journal of Electronics & Information Technology ， 1996 ， 22 （ S1 ）： 41 - 47 .

李锋，周新伦 . 甲骨文自动识别的图论方法［J］. 电子与信息学报， 1996 ， 22 （ S1 ）： 41 - 47 .

LYU X Q ， LI M N ， CAI K W ， et al . An oracle bone inscriptions classification method based on graphic recognition ［J］. Journal of Beijing Information Science & Technology University （Natural Science Edition）， 2010 ， 25 （ S2 ）： 92 - 96 .

吕肖庆，李沫楠，蔡凯伟，等 . 一种基于图形识别的甲骨文分类方法［J］. 北京信息科技大学学报（自然科学版）， 2010 ， 25 （ S2 ）： 92 - 96 .

MENG L . Two-stage recognition for oracle bone inscriptions ［C］// International Conference on Image Analysis and Processing ， 2017 ： 672 - 682 . doi: 10.1007/978-3-319-68548-9_61 http://dx.doi.org/10.1007/978-3-319-68548-9_61

SIMONYAN K ， ZISSERMAN A . Very deep convolutional networks for large-scale image recognition［DB/OL］ . https：//arxiv.org/abs/1409.1556 https://arxiv.org/abs/1409.1556 .

HE K ， ZHANG X ， REN S ， et al . Deep residual learning for image recognition ［C］// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition ， 2016 ： 770 - 778 . doi: 10.1109/cvpr.2016.90 http://dx.doi.org/10.1109/cvpr.2016.90

ZHANG Y K ， ZHANG H ， LIU Y G ， et al . Oracle character recognition based on cross-modal deep metric learning ［J］. Acta Automatica Sinica ， 2021 ， 47 （ 4 ）： 791 - 800 . doi: 10.16383/j.aas.c200443 http://dx.doi.org/10.16383/j.aas.c200443

张颐康，张恒，刘永革，等 . 基于跨模态深度度量学习的甲骨文字识别［J］. 自动化学报， 2021 ， 47 （ 4 ）： 791 - 800 . doi: 10.16383/j.aas.c200443 http://dx.doi.org/10.16383/j.aas.c200443

MAO Y F ， BI X J . Improved ResNeSt network for rubbings of oracle-bone character recognition ［J］. Journal of Intelligent Systems ， 2023 ， 18 （ 3 ）： 450 - 458 .

毛亚菲，毕晓君 . 改进 ResNeSt 网络的拓片甲骨文字识别［J］. 智能系统学报， 2023 ， 18 （ 3 ）： 450 - 458 .

ZHOU X L ， LI F ， HUA X C ， et al . Research on the computer recognition method of oracle bone inscriptions ［J］. Journal of Fudan University（Natural Science）， 1996 ， 35 （ 5 ）： 481 - 486 .

周新伦，李锋，华星城，等 . 甲骨文计算机识别方法研究［J］. 复旦学报（自然科学版）， 1996 ， 35 （ 5 ）： 481 - 486 .

WANG M ， DENG W ， HU J ， et al . Unsupervised attention regularization based domain adaptation for oracle character recognition［DB/OL］ . https：//arxiv.org/abs/2409.15893 https://arxiv.org/abs/2409.15893 .

SEDERBERG T W . Free-form deformation of solid geometric models ［J］. Computer Graphics （SIGGRAPH）， 1992 ， 26 ： 151 .

KIRILLOV A ， MINTUN E ， RAVI N ， et al . Segment anything ［C］// Proceedings of the IEEE/CVF International Conference on Computer Vision ， 2023 ： 4015 - 4026 . doi: 10.1109/iccv51070.2023.00371 http://dx.doi.org/10.1109/iccv51070.2023.00371

LI W ， PENG Y ， ZHANG M ， et al . Deep model fusion： a survey［DB/OL］ . https：//arxiv.org/abs/2309.15698 https://arxiv.org/abs/2309.15698 . doi: 10.1109/tnnls.2025.3628666 http://dx.doi.org/10.1109/tnnls.2025.3628666

ZHANG Y K ， ZHANG H ， LIU Y G ， et al . Oracle character recognition by nearest neighbor classification with deep metric learning ［C］// 2019 International Conference on Document Analysis and Recognition ， 2019 ： 309 - 314 . doi: 10.1109/icdar.2019.00057 http://dx.doi.org/10.1109/icdar.2019.00057

WANG P ， ZHANG K ， WANG X ， et al . An open dataset for oracle bone script recognition and decipherment［DB/OL］ . https：//arxiv.org/abs/2401.15365 https://arxiv.org/abs/2401.15365 . doi: 10.1038/s41597-024-03807-x http://dx.doi.org/10.1038/s41597-024-03807-x

HUANG S ， WANG H ， LIU Y ， et al . OBC306： a large-scale oracle bone character recognition dataset ［C］// 2019 International Conference on Document Analysis and Recognition ， 2019 ： 681 - 688 . doi: 10.1109/icdar.2019.00114 http://dx.doi.org/10.1109/icdar.2019.00114

WANG M ， DENG W . A dataset of oracle characters for benchmarking machine learning algorithms ［J］. Scientific Data ， 2024 ， 11 （ 1 ）： 87 . doi: 10.1038/s41597-024-02933-w http://dx.doi.org/10.1038/s41597-024-02933-w

XING J ， LIU G ， XIONG J . Oracle bone inscription detection： a survey of oracle bone inscription detection based on deep learning algorithm ［C］// Proceedings of the International Conference on Artificial Intelligence， Information Processing and Cloud Computing ， 2019 ： 1 - 8 . doi: 10.1145/3371425.3371434 http://dx.doi.org/10.1145/3371425.3371434

GUAN H ， YANG H ， WANG X ， et al . Deciphering oracle bone language with diffusion models［DB/OL］ . https：//arxiv.org/abs/2406.00684 https://arxiv.org/abs/2406.00684 . doi: 10.18653/v1/2024.acl-long.831 http://dx.doi.org/10.18653/v1/2024.acl-long.831

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Human Jumping Knee-Bending Angle Recognition Model Based on Multi-Channel Attention

Sensitive Text Recognition Based on Deep Bidirectional LSTM and MultiHeadAttention

A Review of Deep Learning-Based Adrenal CT Image Segmentation Research

Research on Prediction Method of Soil Available Nitrogen Content in Guangxi Based on SA-CNN-LSTM

Remote Sensing Image Object Detection Method Integrating Segmentation and Classification Models

Related Author

DING Lei

FANG Chen’an

HU Xinrong

LI Yang

MING Wenkai

LIN Lin

WU Yuan

ZHU Dingju

Related Institution

School of Computer Science and Artificial Intelligence， Wuhan Textile University

Department of Physical Education， Wuhan Textile University

School of Artificial Intelligence， South China Normal University

School of Health Science and Engineering，University of Shanghai for Science and Technology

Department of Cardiology，Shanghai General Hospital，Shanghai Jiao Tong University School of Medicine

⁰