
Chinese General Practice ›› 2025, Vol. 28 ›› Issue (09): 1128-1136.DOI: 10.12114/j.issn.1007-9572.2024.0394
Special Issue: 数智医疗最新文章合辑; 神经/行为/心理最新文章合辑
• Original Research • Previous Articles Next Articles
Received:2024-06-10
Revised:2024-10-10
Published:2025-03-20
Online:2025-01-02
Contact:
ZHOU Yanting, CHEN Jian
通讯作者:
周燕婷, 陈健
作者简介:作者贡献:
王甘红、陈健进行文章的构思与设计;奚美娟、夏开建、张子豪进行数据收集及数据整理,并进行统计学处理与代码报错解决;王甘红、周燕婷撰写论文并进行论文的修订;陈健对文章整体负责,监督管理。
基金资助:
Add to citation manager EndNote|Ris|BibTeX
URL: https://www.chinagp.net/EN/10.12114/j.issn.1007-9572.2024.0394
| 项目 | 艾叶 | 阿胶 | 白扁豆 | 百部 | 白矾 | 百合 | 白蔻 | 白茅根 | 白芍 | 白头翁 | 白术 | 柏子仁 | 巴戟天 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 训练集 | 1 634 | 1 557 | 1 567 | 1 713 | 1 567 | 1 565 | 1 553 | 1 543 | 1 493 | 1 565 | 1 571 | 1 567 | 1 617 |
| 验证集 | 58 | 63 | 53 | 69 | 53 | 55 | 67 | 77 | 69 | 55 | 49 | 53 | 48 |
| 测试集 | 70 | 75 | 54 | 58 | 69 | 58 | 75 | 57 | 44 | 61 | 75 | 58 | 56 |
| 合计 | 1 762 | 1 695 | 1 674 | 1 840 | 1 689 | 1 678 | 1 695 | 1 677 | 1 606 | 1 681 | 1 695 | 1 678 | 1 721 |
Table 1 Number of images for TCM herbals
| 项目 | 艾叶 | 阿胶 | 白扁豆 | 百部 | 白矾 | 百合 | 白蔻 | 白茅根 | 白芍 | 白头翁 | 白术 | 柏子仁 | 巴戟天 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 训练集 | 1 634 | 1 557 | 1 567 | 1 713 | 1 567 | 1 565 | 1 553 | 1 543 | 1 493 | 1 565 | 1 571 | 1 567 | 1 617 |
| 验证集 | 58 | 63 | 53 | 69 | 53 | 55 | 67 | 77 | 69 | 55 | 49 | 53 | 48 |
| 测试集 | 70 | 75 | 54 | 58 | 69 | 58 | 75 | 57 | 44 | 61 | 75 | 58 | 56 |
| 合计 | 1 762 | 1 695 | 1 674 | 1 840 | 1 689 | 1 678 | 1 695 | 1 677 | 1 606 | 1 681 | 1 695 | 1 678 | 1 721 |
| 模型 | 准确率(%) | 精确率(%) | 灵敏度(%) | F1分数(%) | AUC |
|---|---|---|---|---|---|
| EfficientNetB0 | 99.04 | 99.02 | 99.04 | 98.99 | 0.994 2 |
| MobileNetV3 | 99.03 | 99.04 | 99.03 | 99.02 | 0.992 7 |
| ResNet50 | 86.71 | 86.44 | 86.75 | 86.36 | 0.878 9 |
| VGG19 | 98.00 | 97.86 | 97.98 | 97.84 | 0.978 6 |
| ResNet18 | 77.60 | 77.10 | 77.27 | 76.79 | 0.779 1 |
Table 2 Comparison of the performance of AI models in the validation dataset
| 模型 | 准确率(%) | 精确率(%) | 灵敏度(%) | F1分数(%) | AUC |
|---|---|---|---|---|---|
| EfficientNetB0 | 99.04 | 99.02 | 99.04 | 98.99 | 0.994 2 |
| MobileNetV3 | 99.03 | 99.04 | 99.03 | 99.02 | 0.992 7 |
| ResNet50 | 86.71 | 86.44 | 86.75 | 86.36 | 0.878 9 |
| VGG19 | 98.00 | 97.86 | 97.98 | 97.84 | 0.978 6 |
| ResNet18 | 77.60 | 77.10 | 77.27 | 76.79 | 0.779 1 |
| 类别 | 精确率 | 灵敏度 | 特异度 | F1分数 | 准确率 | 平均精度 | AUC | 马修斯相关系数 | 科恩卡帕系数 |
|---|---|---|---|---|---|---|---|---|---|
| 艾叶 | 1 | 0.983 | 1 | 0.991 | 0.983 | 1 | 1 | 0.991 | 0.991 |
| 阿胶 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 白扁豆 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 百部 | 0.971 | 0.986 | 1 | 0.978 | 0.986 | 1 | 1 | 0.978 | 0.978 |
| 白矾 | 0.981 | 1 | 1 | 0.991 | 1 | 1 | 1 | 0.991 | 0.991 |
| 百合 | 0.982 | 1 | 1 | 0.991 | 1 | 0.999 | 1 | 0.991 | 0.991 |
| 白花蛇舌草 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 白茅根 | 0.987 | 1 | 1 | 0.994 | 1 | 1 | 1 | 0.994 | 0.993 |
| 白芍 | 1 | 0.986 | 1 | 0.993 | 0.986 | 1 | 1 | 0.993 | 0.993 |
| 麦芽 | 0.510 | 0.754 | 0.995 | 0.608 | 0.754 | 0.564 | 0.997 | 0.617 | 0.605 |
| 牡丹皮 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 牡蛎 | 0.982 | 0.982 | 1 | 0.982 | 0.982 | 1 | 1 | 0.982 | 0.982 |
| 木香 | 1 | 0.970 | 1 | 0.985 | 0.970 | 1 | 1 | 0.985 | 0.985 |
| 牛膝 | 0.968 | 1 | 1 | 0.984 | 1 | 1 | 1 | 0.984 | 0.984 |
| 总体(加权平均) | 0.990 | 0.990 | 1 | 0.989 | 0.990 | 0.994 | 1 | 0.990 | 0.989 |
Table 3 Evaluation of the performance of the EfficientNetB0 model in the test dataset
| 类别 | 精确率 | 灵敏度 | 特异度 | F1分数 | 准确率 | 平均精度 | AUC | 马修斯相关系数 | 科恩卡帕系数 |
|---|---|---|---|---|---|---|---|---|---|
| 艾叶 | 1 | 0.983 | 1 | 0.991 | 0.983 | 1 | 1 | 0.991 | 0.991 |
| 阿胶 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 白扁豆 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 百部 | 0.971 | 0.986 | 1 | 0.978 | 0.986 | 1 | 1 | 0.978 | 0.978 |
| 白矾 | 0.981 | 1 | 1 | 0.991 | 1 | 1 | 1 | 0.991 | 0.991 |
| 百合 | 0.982 | 1 | 1 | 0.991 | 1 | 0.999 | 1 | 0.991 | 0.991 |
| 白花蛇舌草 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 白茅根 | 0.987 | 1 | 1 | 0.994 | 1 | 1 | 1 | 0.994 | 0.993 |
| 白芍 | 1 | 0.986 | 1 | 0.993 | 0.986 | 1 | 1 | 0.993 | 0.993 |
| 麦芽 | 0.510 | 0.754 | 0.995 | 0.608 | 0.754 | 0.564 | 0.997 | 0.617 | 0.605 |
| 牡丹皮 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| 牡蛎 | 0.982 | 0.982 | 1 | 0.982 | 0.982 | 1 | 1 | 0.982 | 0.982 |
| 木香 | 1 | 0.970 | 1 | 0.985 | 0.970 | 1 | 1 | 0.985 | 0.985 |
| 牛膝 | 0.968 | 1 | 1 | 0.984 | 1 | 1 | 1 | 0.984 | 0.984 |
| 总体(加权平均) | 0.990 | 0.990 | 1 | 0.989 | 0.990 | 0.994 | 1 | 0.990 | 0.989 |
| [1] | |
| [2] | |
| [3] |
曹雪晓,任晓亮,王萌,等. 中药材及饮片规格等级质量标准研究进展[J]. 中药材,2021,44(2):490-494. DOI:10.13863/j.issn1001-4454.2021.02.044.
|
| [4] |
黄丽,张军,吴慧玲,等. 基于深度学习的内镜超声胆管扫查辅助分站系统构建[J]. 中华消化内镜杂志,2022,39(4):295-300. DOI:10.3760/cma.j.cn321463-20210628-00007.
|
| [5] |
吴树剑,俞咏梅,范莉芳,等. 基于增强CT深度学习影像组学术前预测胸腺瘤风险分类[J]. 中国肿瘤临床,2023,50(19):999-1005. DOI:10.12354/j.issn.1000-8179.2023.20230828.
|
| [6] |
陈健,王珍妮,夏开建,等. 基于深度学习的结直肠息肉内镜图像分割和分类方法比较[J]. 上海交通大学学报(医学版),2024,44(6):762-772.
|
| [7] |
郭丛,田钰嘉,李杨,等. 基于YOLOv4算法的中药饮片识别[J]. 中国实验方剂学杂志,2023,29(14):133-140. DOI:10.13422/j.cnki.syfjx.20230614.
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
许郭婷,吴爱荣,林嘉希,等. 基于深度卷积神经网络的上消化道内镜解剖分类模型构建[J]. 中国医学物理学杂志,2023,40(8):1051-1056. DOI:10.3969/j.issn.1005-202X.2023.08.021.
|
| [16] |
陈健,张子豪,卢勇达,等. 基于深度学习构建结直肠息肉诊断自动分类模型[J]. 中华诊断学电子杂志,2024,12(1):9-17.
|
| [17] |
王超超,张先超,谷正昌,等. 中药材及饮片检测中人工智能应用探讨[J]. 中国工程科学,2024,26(2):245-254.
|
| [18] |
胡晓东. 基于改进深度学习算法的中药饮片图像识别研究[D]. 长春:吉林农业大学,2023. DOI:10.27163/d.cnki.gjlnu.2023.000042.
|
| [19] |
|
| [20] |
林嘉希,汪盛嘉,赵鑫,等. 基于深度卷积神经网络的Barrett食管内镜图片分类模型的建立[J]. 上海交通大学学报(医学版),2022,42(5):653-659. DOI:10.3969/j.issn.1674-8115.2022.05.014.
|
| [21] |
|
| [1] | TIAN Xiangyang, ZHOU Yingsheng, GAO Ying. Prospects of Behavioral and Lifestyle Interventions for Diabetes Mellitus [J]. Chinese General Practice, 2026, 29(31): 4513-4520. |
| [2] | XIE Yuanling, CHENG Li, LONG Xiangyu, SU Zhengxin, SU Yiyi, ZHANG Renyue, SUN Guoqiang, CHEN Xuxia, ZHOU Jilan. Effectiveness of an AI-powered Ophthalmic Telemedicine Diagnostic Platform in Community Eye Health Screening [J]. Chinese General Practice, 2026, 29(31): 4641-4649. |
| [3] | WEI Xuejuan, WANG Lihua, WANG Li, REN Jinglin, YAN Haiqing, YU Haiyang, LIU Jingfeng. Effectiveness of an Intelligent Voice Outbound Call System as an Adjunct to Usual Care for Health Management in Older Community-dwelling Patients with "Three Highs" Multimorbidity: a Randomized Controlled Trial [J]. Chinese General Practice, 2026, 29(31): 4635-4640. |
| [4] | CHEN Cheng, CHEN Kaiyuan, SUN Nailing, ZENG Huatang, LEI Suwen, LIANG Wannian. Status and Suggestions on Standardization Construction of Artificial Intelligence-enabled Disease Prevention and Control [J]. Chinese General Practice, 2026, 29(29): 4193-4199. |
| [5] | CHEN Yujing, WANG Yan, LI Guang, DONG Benchao, YANG Peichuan, MA Jianxiong. Digital Technologies in the Management of Sarcopenia among Older Adults in Asia: a Scoping Review [J]. Chinese General Practice, 2026, 29(29): 4200-4210. |
| [6] | HUANG Mu, ZHANG Daochun, YANG Wei, ZHONG Liming, YUAN Wenjing, JIA Ziqi, TAN Xiangliang, DUAN Xiaohui, LIU Xian, CHEN Weicui. Automatic Segmentation of Colorectal Cancer Lesions in Abdominal CT Images Using a Deep Learning Network Based on Fused 3D U-Net and Transformer: a Multicenter and Multidevice Study [J]. Chinese General Practice, 2026, 29(29): 4281-4287. |
| [7] | MU Hongyu, LI Qianqian, LIU Lu, SUN Xinru, ZHANG Jiaqi, DUAN Yeqing, LI Bin. Research Progress and Challenges of Cognitive Digital Therapeutics in the Intervention of Mild Cognitive Impairment [J]. Chinese General Practice, 2026, 29(29): 4211-4216. |
| [8] | XIE Xiaoyun, LIU Chenxi, LIU Yushu, ZHOU Yihan. Artificial Intelligence Research Practices in Primary Healthcare Institutions in China and the Potential Issue of Data Poverty: a Scoping Review [J]. Chinese General Practice, 2026, 29(28): 4067-4077. |
| [9] | DENG Wei, HUANG Mengting, YI Shengli, LIN Yang, XU Lidan, REN Zhexue, ZHU Xiuqiong, LI Yesha, REN Mengge, KE Dazhi, LI Guiqiong, WU Zhiqin, LUO Aoran, HE Fapei. Training and Effectiveness of AI-enabled Clinical Reasoning in General Practice [J]. Chinese General Practice, 2026, 29(28): 4078-4085. |
| [10] | ZHOU Yuhan, MIN Yuqi, GUO Zhou, YANG Min, YUE Lu, HE Meichen, ZHOU Hairong, WANG Yongchen. Opportunities and Challenges for General Practitioners in the Age of Artificial Intelligence [J]. Chinese General Practice, 2026, 29(25): 3563-3568. |
| [11] | WANG Hongchuan, ZHANG Jianbo, MA Wei, ZHAO Sidi. Governance Logic and Ecosystem Construction of Healthcare Data Space Led by Digital Policies: an Analytical Framework Based on "Technology-Management-Law" [J]. Chinese General Practice, 2026, 29(24): 3416-3422. |
| [12] | OUYANG Weina, WANG Lu, ZHANG Xuan, TANG Wei, FAN Yongmei. Research on Atrial Fibrillation Screening Based on Dynamic Multi-scale Feature Fusion and Attention Mechanism [J]. Chinese General Practice, 2026, 29(23): 3281-3286. |
| [13] | CHEN Kaiyuan, WANG Hongchuan, GE Yi, SUN Nailing, CHEN Cheng, WANG Tianran, ZENG Huatang, LEI Suwen, LIANG Wannian. The Current Situation and Challenges of Building a Standards System for High-quality Development of Health Care Empowered by Generative Artificial Intelligence [J]. Chinese General Practice, 2026, 29(22): 3073-3078. |
| [14] | WANG Yang, ZOU Chuan, GUO Mengruo, HE Liuhua, LU Xuhang, GUI Xianzhong, XU Zhijie, LIN Kai, JIN Hua, YAO Mi, YANG Hui, YU Dehua. A Five-stage Approach for Primary Care Practitioners to Refine High-quality Research Questions with Artificial Intelligence Assistance [J]. Chinese General Practice, 2026, 29(22): 3094-3102. |
| [15] | HU Min, LYU Xiangdong. Accuracy of Artificial Intelligence in Remote Electrocardiography Diagnosis [J]. Chinese General Practice, 2026, 29(18): 2498-2503. |
| Viewed | ||||||
|
Full text |
|
|||||
|
Abstract |
|
|||||