《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (4): 1325-1332.DOI: 10.11772/j.issn.1001-9081.2024040438

• 多媒体计算与计算机仿真 • 上一篇    下一篇

多尺度2D-Adaboost的中药材粉末显微图像识别算法

王一丁1, 王泽浩1, 李耀利2, 蔡少青2(), 袁媛3   

  1. 1.北方工业大学 信息学院,北京 100144
    2.北京大学 药学院,北京 100191
    3.中国中医科学院 中药资源中心,北京 100700
  • 收稿日期:2024-04-16 修回日期:2024-06-18 接受日期:2024-06-26 发布日期:2025-04-08 出版日期:2025-04-10
  • 通讯作者: 蔡少青
  • 作者简介:王一丁(1967—),男,北京人,教授,博士,CCF会员,主要研究方向:生物特征识别、机器视觉
    王泽浩(1999—),男,北京人,硕士研究生,主要研究方向:计算机视觉、深度学习
    李耀利(1977—),女,陕西宝鸡人,讲师,博士,主要研究方向:中药鉴定和质量评价
    袁媛(1978—),女,河北保定人,研究员,博士,主要研究方向:中药鉴定、分子生药学。
  • 基金资助:
    中央本级重大增减支项目(2060302)

Multi-scale 2D-Adaboost microscopic image recognition algorithm of Chinese medicinal materials powder

Yiding WANG1, Zehao WANG1, Yaoli LI2, Shaoqing CAI2(), Yuan YUAN3   

  1. 1.School of Information Science and Technology,North China University of Technology,Beijing 100144,China
    2.School of Pharmaceutical Sciences,Peking University,Beijing 100191,China
    3.National Resource Center for Chinese Materia Medica,China Academy of Chinese Medical Sciences,Beijing 100700,China
  • Received:2024-04-16 Revised:2024-06-18 Accepted:2024-06-26 Online:2025-04-08 Published:2025-04-10
  • Contact: Shaoqing CAI
  • About author:WANG Yiding, born in 1967, Ph. D., professor. His research interests include biometrics recognition, machine vision.
    WANG Zehao, born in 1999, M. S. candidate. His research interests include computer vision, deep learning.
    LI Yaoli, born in 1977, Ph. D., lecturer. Her research interests include identification and quality evaluation of traditional Chinese medicines.
    YUAN Yuan, born in 1978, Ph. D., research fellow. Her research interests include identification of traditional Chinese medicines, molecular pharmacognosy.
  • Supported by:
    Key Project at the Central Government Level(2060302)

摘要:

针对中药材粉末的显微图像中含有大量细微特征和背景干扰因素导致的同一类药材的变化过大(类内差异大)和多种药材之间特征过于相似(类间差异小)的问题,提出一种多尺度2D-Adaboost算法。首先,构建一个全局?局部特征融合的主干网络架构,以更好地提取多尺度特征,该架构通过结合Transformer和卷积神经网络(CNN)的优势能有效提取并融合各个尺度的全局和局部特征,从而显著提高主干网络的特征捕捉能力;其次,将Adaboost的单尺度输出拓展到多尺度,并构建2D-Adaboost结构的背景抑制模块,该模块将主干网络各个尺度的输出特征图划分为前景和背景,从而有效抑制背景区域的特征值,并增加判别性特征的强度;最后,在2D-Adaboost结构的每个尺度上额外添加一个分类器以构建特征细化模块,该模块通过控制温度参数协调分类器间的协作学习,从而逐步细化不同尺度的特征图,帮助网络学习更合适的特征尺度,并丰富细节特征的表示。实验结果表明,所提算法的识别准确率达到了96.85%,与ConvNeXt-L、ViT-L、Swin-L和Conformer-L模型相比分别上升了7.56、5.26、3.79和2.60个百分点。高准确率和分类效果的稳定性验证了所提算法在中药材粉末显微图像分类任务中的有效性。

关键词: 深度学习, 中药材, 显微图像识别, 特征融合, 2D-Adaboost

Abstract:

A multi-scale 2D-Adaboost algorithm was proposed to solve the problem that the microscopic images of Chinese medicinal materials powder contain a large number of fine features and background interference factors, which leads to excessive changes in the same medicinal materials (large differences within the class) and too similar features among various medicinal materials (small differences between the classes). Firstly, a global-local feature fusion backbone network architecture was constructed to extract multi-scale features better. By combining the advantages of Transformer and Convolutional Neural Network (CNN), this architecture was able to extract and fuse global and local features at various scales effectively, thereby improving the feature capture capability of the backbone network significantly. Secondly, the single-scale output of Adaboost was extended to multi-scale output, and a 2D-Adaboost structure-based background suppression module was constructed. With this module, the output feature maps of each scale of the backbone network were divided into foreground and background, thereby suppressing feature values of the background region effectively and enhancing the strength of discriminative features. Finally, an extra classifier was added to each scale of the 2D-Adaboost structure to build a feature refinement module, which coordinated the collaborative learning among the classifiers by controlling temperature parameters, thereby refining the feature maps of different scales gradually, helping the network to learn more appropriate feature scales, and enriching the detailed feature representation. Experimental results show that the recognition accuracy of the proposed algorithm reaches 96.85%, which is increased by 7.56, 5.26, 3.79 and 2.60 percentage points, respectively, compared with those of ConvNeXt-L, ViT-L, Swin-L, and Conformer-L models. The high accuracy and stability of the classification validate the effectiveness of the proposed algorithm in classification tasks of Chinese medicinal materials powder microscopic images.

Key words: deep learning, Chinese medicinal material, microscopic image recognition, feature fusion, 2D-Adaboost

中图分类号: