基于频域特征与多任务学习的中国古钱币包浆鉴别方法

doi:10.16451/j.cnki.issn1003-6059.202607004

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (4506 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要针对古钱币真伪包浆宏观视觉高度相似导致的传统空域模型判别能力不足,以及现有方法难以有效解决包浆物理演化程度的连续评分问题,提出基于频域特征与多任务学习的中国古钱币包浆鉴别方法.首先,设计频域特征提取与增强模块,通过分块离散余弦变换捕捉真伪包浆在频域能量分布上的差异,利用双路池化策略感知异常频率.然后,构建多尺度空频调制融合模块,将频域特征转化为注意力权重,调制空域特征,避免语义冲突.最后,设计融合特征筛选模块,抑制钱文轮廓干扰.同时构建多任务联合学习框架以求解评分,实现真伪判别与包浆评分连续回归的联合优化.在自建数据集上的实验表明,文中方法在真伪判别任务与包浆评分任务上均取得较优性能.

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	黄捷
	田展宏

关键词 ：古钱币, 包浆鉴别, 频域特征, 深度学习, 特征融合

Abstract：The high macroscopic visual similarity between authentic and fake patina on ancient coins limits the discriminative capability of traditional spatial-domain models. Existing methods struggle to solve the continuous scoring problem of the physical evolution degree of patina. To address the above issues, a Chinese ancient coin patina authentication method based on frequency-domain features and multi-task learning(FDF-MTL) is proposed. First, a frequency-domain feature extraction and enhancement module is designed. Patch-based discrete cosine transform is adopted to capture the essential differences in frequency energy distribution between authentic patina and fake patina. A dual-path pooling strategy is utilized to perceive abnormal frequencies. Second, a multi-scale spatial-frequency modulation fusion module is constructed to transform frequency-domain features into attention weights for modulating spatial-domain features, thereby avoiding semantic conflicts. Finally, a fused feature screening module is designed to suppress the interference of coin inscription contours, and a multi-task learning framework is constructed to solve the scoring problem. The joint optimization of authenticity discrimination and continuous patina score regression is achieved. Experiments on a self-built dataset demonstrate the superior performance of the proposed method in both authenticity discrimination and patina scoring tasks.

Key words： Ancient Coin Patina Authentication Frequency-Domain Feature Deep Learning Feature Fusion

收稿日期: 2026-02-06

ZTFLH:

TP 391

通讯作者: 黄捷,博士,教授,主要研究方向为模式识别、智能系统、多智能体系统.E-mail:jie.huang@fzu.edu.cn.

作者简介: 田展宏,硕士研究生,主要研究方向为模式识别.E-mail:zhanhongtian@fzu.edu.cn.

引用本文:

黄捷, 田展宏. 基于频域特征与多任务学习的中国古钱币包浆鉴别方法[J]. 模式识别与人工智能, 2026, 39(7): 621-635. HUANG Jie, TIAN Zhanhong. Chinese Ancient Coin Patina Authentication Method Based on Frequency-Domain Features and Multi-task Learning. Pattern Recognition and Artificial Intelligence, 2026, 39(7): 621-635.

链接本文:

http://manu46.magtech.com.cn/Jweb_prai/CN/10.16451/j.cnki.issn1003-6059.202607004 或 http://manu46.magtech.com.cn/Jweb_prai/CN/Y2026/V39/I7/621

[1] 陈浩标. 北宋古钱币的鉴定方法与要则[J].收藏与投资, 2025, 16(12): 126-128.
(Chen H B. Authentication methods and key principles of northern song dynasty coins[J]. Collection and Investment, 2025, 16(12): 126-128.)
[2] 周卫荣. 中国古代钱币与钱币科技考古[J].南方文物, 2007(1): 9-14.
(Zhou W R. Chinese ancient coins and scientific archaeology of coins[J].Cultural Relics in Southern China, 2007(1): 9-14.)
[3] Zaharieva M, Huber-Mörk R, Nölle M, et al. On ancient coin classi-fication[C]//Proceedings of the 8th International Symposium on Virtual Reality, Archaeology and Cultural Heritage. Eindhoven, The Netherlands: Eurographics Association, 2007: 55-62.
[4] Zambanini S, Kampel M. Automatic coin classification by image matching[C]//Proceedings of the 12th International Conference on Virtual Reality, Archaeology and Cultural Heritage. Eindhoven, The Netherlands: Eurographics Association, 2011: 65-72.
[5] Schlag I, Arandjelović O. Ancient Roman coin recognition in the wild using deep learning based recognition of artistically depicted face profiles[C]// Proceedings of the IEEE International Confe-rence on Computer Vision Workshops. Washington, USA: IEEE, 2017: 2898-2906.
[6] Wei K P, He B, Wang F, et al. A novel method for classification of ancient coins based on image textures[C]//Proceedings of the 2nd Workshop on Digital Media and Its Application in Museum and He-ritages. Washington, USA: IEEE, 2007: 63-66.
[7] Hui J, Mi W Y, Wang J J, et al. AFSIFormer: adaptive frequency-spatial interaction attention mechanism for aerial image semantic segmentation[J/OL]. IEEE Transactions on Geoscience and Remote Sensing, 2025, 63. https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11126528.
[8] 赵燕生. 中国钱币鉴定评级市场发展报告[J].中国钱币, 2016(5): 64-80.
(Zhao Y S. Development report of chinese coin authentication and grading market[J]. China Numismatics, 2016(5): 64-80.)
[9] Aslan S, Vascon S, Pelillo M. Two sides of the same coin: improved ancient coin classification using graph transduction games. Pattern Recognition Letters, 2020, 131: 158-165.
[10] Lin W J, Jhuo S S. Coin recognition based on texture classification on ring and fan areas of the coin image[C/OL]//Proceedings of the IEEE International Instrumentation and Measurement Technology Conference. Washington, USA: IEEE, 2016. https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7520507&tag=1.
[11] Arandjelović O. Reading ancient coins: automatically identifying de-narii using obverse legend seeded retrieval[C]//Proceedings of the 12th European Conference on Computer Vision. Berlin, Germany: Springer, 2012: 317-330.
[12] Kavelar A, Zambanini S, Kampel M. Word detectionapplied to images of ancient Roman coins[C]//Proceedings of the 18th International Conference on Virtual Systems and Multimedia. Washington, USA: IEEE, 2012: 577-580.
[13] Kim J, Pavlovic V. Improving ancient Roman coin recognition with alignment and spatial encoding[C]//Proceedings of the 13th European Conference on Computer Vision. Berlin, Germany: Springer, 2015: 149-164.
[14] Kim J, Pavlovic V. Discovering characteristic landmarks on ancient coins using convolutional networks[C]//Proceedings of the 23rd International Conference on Pattern Recognition. Washington, USA: IEEE, 2016: 1595-1600.
[15] 李耀明,杨延喆.基于卷积神经网络的钱币图像识别方法研究[J].无线互联科技, 2024, 21(4): 1-5.
(Li Y M, Yang Y Z. Research on coin recognition method based on convolutional neural networks[J]. Wireless Internet Science and Technology, 2024, 21(4): 1-5.)
[16] 施雨. 基于神经网络的古钱币图像识别检索研究[D].硕士学位论文.淮南:安徽理工大学, 2020.
(Shi Y. Research on image recognition and retrieval of ancient coins based on neural network[D]. Master Dissertation. Huainan, China: Anhui University of Science and Technology, 2020.)
[17] Cirillo S, Solimando G, Virgili L. A deep learning approach to classify country and value of modern coins[J]. Neural Computing and Applications, 2024, 36(20): 11759-11775.
[18] Bassett R, Gallivan P, Gao X, et al. Development of an automated coin grader: a progress report[C]//Proceedings of the 8th Mid-Atlantic Student Workshop on Programming Languages and Systems. New York, USA: Pace University Press, 2002: 15.1-15.10.
[19] Vadivelan A, Praveen K. Implementation of coin recognition by extracting the local texture features with gabor wavelet and LBP ope-rator[EB/OL]. [2026-01-05].https://www.academia.edu/download/31593331/135-139.pdf.
[20] Pan X Y, Tougne L. Image analysis and deep learning for aiding professional coin grading[EB/OL]. [2026-01-05].https://hal.science/hal-01795304/document.
[21] Gueguen L, Sergeev A, Kadlec B, et al. Faster neural networks straight from JPEG[C]//Proceedings of the 32nd International Conference on Neural Information Processing Systems. Cambridge, USA: MIT Press, 2018: 3937-3948.
[22] Ehrlich M, Davis L S. Deep residual learning in the JPEG transform domain[C]//Proceedings of the IEEE/CVF International Confe-rence on Computer Vision. Washington, USA: IEEE, 2019: 3483-3492.
[23] Xu K, Qin M H, Sun F, et al. Learning in the frequency domain[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2020: 1737-1746.
[24] Ahmed N, Natarajan T, Rao K R. Discrete cosine transform[J]. IEEE Transactions on Computers, 1974, C-23(1): 90-93.
[25] Zhong Y J, Li B, Tang L, et al. Detecting camouflaged object in frequency domain[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2022: 4494-4503.
[26] 葛斌,邵孜绎,郑军帅,等.频域增强自适应通道注意力与特征金字塔融合的可逆流网络风格迁移[J].模式识别与人工智能, 2026, 39(1): 83-96.
(Ge B, Shao Z Y, Zheng J S, et al. Reversible flow network for style transfer based on frequency-domain enhanced adaptive channel attention and feature pyramid fusion[J]. Pattern Recognition and Artificial Intelligence, 2026, 39(1): 83-96.)
[27] 林力伟,李阳,朱恒亮,等.结合多级特征解耦与频域信息的深度伪造检测方法[J].模式识别与人工智能, 2026, 39(2): 127-140.
(Lin L W, Li Y, Zhu H L, et al. Deepfake detection method combining multi-stage feature disentanglement and frequency-domain information[J]. Pattern Recognition and Artificial Intelligence, 2026, 39(2): 127-140.)
[28] Mehta S, Rastegari M. MobileViT: light-weight, general-purpose, and mobile-friendly vision transformer[EB/OL]. [2026-01-05].https://arxiv.org/pdf/2110.02178.
[29] Woo S, Park J, Lee J Y, et al. CBAM: convolutional block attention module[C]//Proceedings of the 15th European Conference on Computer Vision. Berlin, Germany: Springer, 2018: 3-19.
[30] 王舒乙. 古钱的坑口与鉴定[J].文物鉴定与鉴赏, 2014(2): 34-39.
(Wang S Y. Kengkou and authentication of ancient coins[J]. Identification and Appreciation to Cultural Relics, 2014(2): 34-39.)
[31] Kiourt C, Evangelidis V. AnCoins: image-based automated identification of ancient coins through transfer learning approaches[C]//Proceedings of the International Workshops and Challenges on Pa-ttern Recognition. Berlin, Germany: Springer, 2021: 54-67.
[32] Anwar H, Anwar S, Zambanini S,et al. Deep ancient Roman republican coin classification via feature fusion and attention. Pattern Recognition, 2021, 114. https://doi.org/10.1016/j.patcog.2021.107871.
[33] Howard A, Sandler M, Chu G, et al. Searching for MobileNetV3[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Washington, USA: IEEE, 2019: 1314-1324.
[34] He K M, Zhang X Y, Ren S Q, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2016: 770-778.
[35] Wang A, Chen H, Lin Z J, et al. RepViT: revisiting mobile CNN from ViT perspective[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2024: 15909-15920.
[36] Qin Z Q, Zhang P Y, Wu F, et al. FcaNet: frequency channel attention networks[C]//Proceedings of the IEEE/CVF Internatio-nal Conference on Computer Vision. Washington, USA: IEEE, 2021: 763-772.
[37] Rao Y M, Zhao W L, Zhu Z, et al. Global filter networks for image classification[C]//Proceedings of the 35th International Conference on Neural Information Processing Systems. Cambridge, USA: MIT Press, 2021: 980-993.
[38] Wang Y, Li Y Y, Wu R Z, et al. CFormer: cross-frequency transformer for remote sensing semantic segmentation[J/OL]. IEEE Transactions on Geoscience and Remote Sensing, 2026, 64. https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11432949.
[39] Kang L, Ye P, Li Y, et al. Convolutional neural networks for no-reference image quality assessment[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2014: 1733-1740.
[40] Golestaneh S A, Dadsetan S, Kitani K M. No-reference image quality assessment via transformers, relative ranking, and self-consistency[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. Washington, USA: IEEE, 2022: 3989-3999.
[41] Yang S D, Wu T H, Shi S W, et al. MANIQA: Multi-dimension attention network for no-reference image quality assessment[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Washington, USA: IEEE, 2022: 1190-1199.
[42] Selvaraju R R, Cogswell M, Das A, et al. Grad-CAM: visual ex-planations from deep networks via gradient-based localization[C]//Proceedings of the IEEE International Conference on Computer Vision. Washington, USA: IEEE, 2017: 618-626.