Dynamic Temporal-Guided Diffusion Model for Pathological Image Classification

ZHAO Heyang; CHEN Ruhao; LUO Yikai; ZHANG Peng; WANG Wenjian

doi:10.12263/DZXB.20251079

您当前的位置：

首页 >

文章列表页 >

Dynamic Temporal-Guided Diffusion Model for Pathological Image Classification

PAPERS | 更新时间：2026-06-17

- Dynamic Temporal-Guided Diffusion Model for Pathological Image Classification
- ACTA ELECTRONICA SINICA Vol. 54, Issue 4, Pages: 1736-1749(2026)
- 作者机构：
  
  1.山西大学计算机与信息技术学院，山西太原 030006
  2.山西大学计算智能与中文信息处理教育部重点实验室，山西太原 030006
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(62506220;62476157;U21A20513);Fundamental Research Program of Shanxi Province(202403021222024);Special Project for Patent Transformation in Shanxi Province(20250016);Scientific and Technological Innovation Programs of Higher Education Institutions in Shanxi(2024L002)
- DOI：10.12263/DZXB.20251079
  CLC： R361.3;TP391.41
- Received：21 January 2026，
  
  Accepted：25 March 2026，
  
  Published：25 April 2026
- 稿件说明：
移动端阅览
赵赫阳, 陈如豪, 罗一恺, 等. 动态时序引导的病理图像分类扩散模型[J]. 电子学报, 2026, 54(04): 1736-1749.

ZHAO Heyang, CHEN Ruhao, LUO Yikai, et al. Dynamic Temporal-Guided Diffusion Model for Pathological Image Classification[J]. Acta Electronica Sinica, 2026, 54(04): 1736-1749.
赵赫阳, 陈如豪, 罗一恺, 等. 动态时序引导的病理图像分类扩散模型[J]. 电子学报, 2026, 54(04): 1736-1749. DOI：10.12263/DZXB.20251079

ZHAO Heyang, CHEN Ruhao, LUO Yikai, et al. Dynamic Temporal-Guided Diffusion Model for Pathological Image Classification[J]. Acta Electronica Sinica, 2026, 54(04): 1736-1749. DOI：10.12263/DZXB.20251079

摘要

基于深度学习的病理图像分类是实现高效、可靠病理辅助诊断的关键技术基础，对病理图像智能分析、病变区域精准识别与组织类型自动判别具有重要研究意义。当前主流病理图像分类方法多依赖单一时间点的静态特征提取，难以有效表征病理状态从正常、发育异常到恶性转化的连续演进过程，无法充分建模组织层面复杂的层级结构与空间分布特征，导致对相似病理区域、细微病变结构的识别能力存在明显局限，模型判别精度与泛化性能难以进一步提升。针对上述问题，本文提出一种动态时序引导的病理图像分类扩散模型（Dynamic Temporal-guided Diffusion for Pathology， DT-DPath），构建统一的病理图像分类框架。该模型以扩散模型为基础，将前向加噪与反向去噪过程模拟为组织形态学的

退化与重构过程，通过聚合扩散过程中不同噪声水平下的时序动态特征，构建能够刻画病理演化过程的判别性表征，精准捕捉细胞层面的细微结构差异，并有效建模组织内部的复杂层级与空间分布关系。在特征提取阶段，本文设计时序动态通道注意力机制（Temporal Dynamic Channel Attention，TDCA），将通道注意力扩展至时序维度，自适应聚焦于关键时序阶段的特征响应，增强模型对时序演化信息的表征能力。为优化训练效率，本文采用基于时序贡献的非均匀重要性采样策略，依据各时间步的分类显著性与信息密度动态构建采样分布，优先保留高判别力的中间去噪阶段，在保持分类性能的同时显著提升模型训练速度与稳定性。实验在结直肠癌、乳腺癌及淋巴结转移等多种病理图像公开数据集上开展，与CNN（Convolutional Neural Networks）、Transformer、扩散模型分类器以及病理基础模型等当前先进方法对比，结果表明DT-DPath在分类准确率、精确率、召回率及

分数等指标上全面领先，性能提升了0.65~3.59个百分点。消融实验验证了时序动态通道注意力与非均匀重要性采样的有效性，可视化分析显示模型可精准聚焦病理判别关键区域，特征聚类效果显著优于传统方法。本文方法实现了从静态特征建模到动态时序演化表征的范式转变，为病理图像智能分类提供了新的技术路径，也为扩散模型在病理图像分析领域的应用提供了可解释的研究思路。

Abstract

Deep learning-based pathological image classification is a key technical foundation for efficient and reliable computer-aided pathological analysis

which is of great significance for intelligent analysis of pathological images

accurate identification of lesion regions and automatic discrimination of tissue types. Current mainstream pathological image classification methods mostly rely on static feature extraction at a single time point

which makes it difficult to effectively represent the continuous evolution process of pathological states from normal

developmental abnormalities to malignant transformation

and cannot fully model the complex hierarchical structure and spatial distribution features at the tissue level. As a result

the recognition ability for similar pathological regions and subtle lesion structures is obviously limited

and the discrimination accuracy and generalization performance of the model are difficult to further improve. To address the above issues

this paper proposes a dynamic temporal-guided diffusion model for pathology (DT-DPath)

which constructs a unified classification framework for pathological images. Based on the diffusion model

this framework simulates the forward noise addition and reverse denoising processes as the degradation and reconstruction of tissue morphology. By

aggregating temporal dynamic features at different noise levels during the diffusion process

it constructs discriminative representations that depict the pathological evolution process

accurately captures subtle structural differences at the cellular level

and effectively models the complex hierarchical and spatial distribution relationships within tissues. In the feature extraction stage

this paper designs a temporal dynamic channel attention (TDCA) mechanism

which extends channel attention to the temporal dimension and adaptively focuses on feature responses at key temporal stages to enhance the model’s ability to represent temporal evolution information. To optimize training efficiency

a non-uniform importance sampling strategy based on temporal contribution is adopted

which dynamically constructs the sampling distribution according to the classification saliency and information density of each time step

and preferentially retains intermediate denoising stages with high discriminative power. This strategy significantly improves training speed and stability while maintaining classification performance. Experiments are conducted on a variety of public pathological image datasets including colorectal cancer

breast cancer and lymph node metastasis. Compared with state-of-the-art methods such as CNN

Transformer

diffusion model classifiers and pathological foundation models

the results show that DT-DPath achieves comprehensive superiority in accuracy

precision

recall and

-score

with performance improvements ranging from 0.65~3.59 percentage points. Ablation experiments verify the effectiveness of temporal dynamic channel attention and non-uniform importance sampling

and visualization analysis shows that the model can accurately focus on key regions for pathological discrimination

with significantly better feature clustering effects than traditional methods. The proposed method realizes the paradigm shift from static feature modeling to dynamic temporal evolution represe

ntation

provides a new technical path for intelligent classification of pathological images

and also offers an interpretable research idea for the application of diffusion models in the field of pathological image analysis.

关键词

Keywords

references

Biller L H , Schrag D . Diagnosis and treatment of metastatic colorectal cancer: A review [J ] . JAMA , 2021 , 325 ( 7 ): 669 - 685 . DOI: 10.1001/jama.2021.0106 http://dx.doi.org/10.1001/jama.2021.0106

Sawicki T , Ruszkowska M , Danielewicz A , et al . A review of colorectal cancer in terms of epidemiology, risk factors, development, symptoms and diagnosis [J ] . Cancers , 2021 , 13 ( 9 ): 2025 . DOI: 10.3390/cancers13092025 http://dx.doi.org/10.3390/cancers13092025

Center M M , Jemal A , Smith R A , et al . Worldwide variations in colorectal cancer [J ] . CA : A Cancer Journal for Clinicians , 2009 , 59 ( 6 ): 366 - 378 . DOI: 10.3322/caac.20038 http://dx.doi.org/10.3322/caac.20038

Boyle P , Leon M E . Epidemiology of colorectal cancer [J ] . British Medical Bulletin , 2002 , 64 ( 1 ): 1 - 25 . DOI: 10.1093/bmb/64.1.1 http://dx.doi.org/10.1093/bmb/64.1.1

Mármol I , Sánchez-de-Diego C , Pradilla Dieste A , et al . Colorectal carcinoma: A general overview and future perspectives in colorectal cancer [J ] . International Journal of Molecular Sciences , 2017 , 18 ( 1 ): 197 . DOI: 10.3390/ijms18010197 http://dx.doi.org/10.3390/ijms18010197

Dekker E , Tanis P J , Vleugels J L , et al . Colorectal cancer [J ] . The Lancet , 2019 , 394 ( 10207 ): 1467 - 1480 . DOI: 10.1016/j.cgh.2019.02.026 http://dx.doi.org/10.1016/j.cgh.2019.02.026

Sarvamangala D R , Kulkarni R V . Convolutional neural networks in medical image understanding: A survey [J ] . Evolutionary Intelligence , 2022 , 15 ( 1 ): 1 - 22 . DOI: 10.1007/s12065-020-00540-3 http://dx.doi.org/10.1007/s12065-020-00540-3

Peng Chuncheng , Lee B R . Enhancing colorectal cancer histological image classification using transfer learning and ResNet50 CNN model [C ] // Proceedings of 2023 IEEE 5th Eurasia Conference on Biomedical Engineering, Healthcare and Sustainability (ECBIOS) . Piscataway : IEEE , 2023 : 36 - 40 . DOI: 10.1109/ecbios57802.2023.10218590 http://dx.doi.org/10.1109/ecbios57802.2023.10218590

Khayatian D , Maleki A , Nasiri H , et al . Histopathology image analysis for gastric cancer detection: A hybrid deep learning and catboost approach [J ] . Multimedia Tools and Applications , 2025 , 84 ( 19 ): 21777 - 21803 . DOI: 10.1007/s11042-024-19816-2 http://dx.doi.org/10.1007/s11042-024-19816-2

Ben Hamida A , Devanne M , Weber J , et al . Weakly supervised learning using attention gates for colon cancer histopathological image segmentation [J ] . Artificial Intelligence in Medicine , 2022 , 133 : 102407 . DOI: 10.1016/j.artmed.2022.102407 http://dx.doi.org/10.1016/j.artmed.2022.102407

Murtaza G , Shuib L , Mujtaba G , et al . Breast cancer multi-classification through deep neural network and hierarchical classification approach [J ] . Multimedia Tools and Applications , 2020 , 79 ( 21 ): 15481 - 15511 . DOI: 10.1007/s11042-019-7525-4 http://dx.doi.org/10.1007/s11042-019-7525-4

He Kelei , Gan Chen , Li Zhuoyuan , et al . Transformers in medical image analysis [J ] . Intelligent Medicine , 2023 , 3 ( 1 ): 59 - 78 . DOI: 10.1016/j.imed.2022.07.002 http://dx.doi.org/10.1016/j.imed.2022.07.002

Shamshad F , Khan S , Zamir S W , et al . Transformers in medical imaging: A survey [J ] . Medical Image Analysis , 2023 , 88 : 102802 . DOI: 10.1016/j.media.2023.102802 http://dx.doi.org/10.1016/j.media.2023.102802

李广丽 , 叶艺源 , 吴光庭 , 等 . 联合多视角Transformer编码与在线融合互学习的乳腺癌病理图像分类模型 [J ] . 电子学报 , 2024 , 52 ( 7 ): 2369 - 2381 . DOI: 10.12263/DZXB.20230305 http://dx.doi.org/10.12263/DZXB.20230305

Li Guangli , Ye Yiyuan , Wu Guangting , et al . Breast cancer pathological image classification model via combining multi-view transformer coding and online fusion mutual learning [J ] . Acta Electronica Sinica , 2024 , 52 ( 7 ): 2369 - 2381 . (in Chinese) . DOI: 10.12263/DZXB.20230305 http://dx.doi.org/10.12263/DZXB.20230305

宋艳涛 , 路云里 . SwinT-Unet: 基于双通道自注意力机制的超声图像分割方法 [J ] . 电子学报 , 2024 , 52 ( 11 ): 3835 - 3846 . DOI: 10.12263/DZXB.20230904 http://dx.doi.org/10.12263/DZXB.20230904

Song Yantao , Lu Yunli . SwinT-Unet: Ultrasound image segmentation based on two-channel self-attention mechanism [J ] . Acta Electronica Sinica , 2024 , 52 ( 11 ): 3835 - 3846 . (in Chinese) . DOI: 10.12263/DZXB.20230904 http://dx.doi.org/10.12263/DZXB.20230904

Kolla B , Venugopal P . An integrated approach for magnification independent breast cancer classification [J ] . Biomedical Signal Processing and Control , 2024 , 88 : 105594 . DOI: 10.1016/j.bspc.2023.105594 http://dx.doi.org/10.1016/j.bspc.2023.105594

Liu M X , Liu Y Z , Xu P B , et al . Exploiting geometric features via hierarchical graph pyramid transformer for cancer diagnosis using histopathological images [J ] . IEEE Transactions on Medical Imaging , 2024 , 43 ( 8 ): 2888 - 2900 . DOI: 10.1109/tmi.2024.3381994 http://dx.doi.org/10.1109/tmi.2024.3381994

Ding Meidan , Qu Aiping , Zhong Haiqin , et al . A transformer-based network for pathology image classification [C ] // Proceedings of 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) . Piscataway : IEEE , 2021 : 2028 - 2034 . DOI: 10.1109/bibm52615.2021.9669476 http://dx.doi.org/10.1109/bibm52615.2021.9669476

Kather J N , Krisam J , Charoentong P , et al . Predicting survival from colorectal cancer histology slides using deep learning: A retrospective multicenter study [J ] . PLoS Medicine , 2019 , 16 ( 1 ): e1002730 . DOI: 10.1371/journal.pmed.1002730 http://dx.doi.org/10.1371/journal.pmed.1002730

Spanhol F A , Oliveira L S , Petitjean C , et al . A dataset for breast cancer histopathological image classification [J ] . IEEE Transactions on Biomedical Engineering , 2016 , 63 ( 7 ): 1455 - 1462 . DOI: 10.1109/tbme.2015.2496264 http://dx.doi.org/10.1109/tbme.2015.2496264

Veeling B S , Linmans J , Winkens J , et al . Rotation equivariant CNNs for digital pathology [C ] // Proceedings of the 21st International Conference on Medical Image Computing and Computer Assisted Intervention . Heidelberg : Springer , 2018 : 210 - 218 . DOI: 10.1007/978-3-030-00934-2_24 http://dx.doi.org/10.1007/978-3-030-00934-2_24

Albashish D , Al-Sayyed R , Abdullah A , et al . Deep CNN model based on VGG16 for breast cancer classification [C ] // Proceedings of 2021 International Conference on Information Technology (ICIT) . Piscataway : IEEE , 2021 : 805 - 810 . DOI: 10.1109/icit52682.2021.9491631 http://dx.doi.org/10.1109/icit52682.2021.9491631

Zhong Ziliang , Zheng Muhang , Huafeng Mai , et al . Cancer image classification based on DenseNet model [J ] . Journal of Physics: Conference Series , 2020 , 1651 ( 1 ): 012143 . DOI: 10.1088/1742-6596/1651/1/012143 http://dx.doi.org/10.1088/1742-6596/1651/1/012143

Gupta M , Verma N , Sharma N , et al . Deep transfer learning hybrid techniques for precision in breast cancer tumor histopathology classification [J ] . Health Information Science and Systems , 2025 , 13 ( 1 ): 20 . DOI: 10.1007/s13755-025-00337-7 http://dx.doi.org/10.1007/s13755-025-00337-7

Dosovitskiy A , Beyer L , Kolesnikov A , et al . An image is worth 16x16 words: Transformers for image recognition at scale [C ] // Proceedings of the 9th International Conference on Learning Representations . [S.l.] : ICLR , 2021 : 1 - 21 .

Touvron H , Cord M , Douze M , et al . Training data-efficient image transformers & distillation through attention [C ] // Proceedings of the 38th International Conference on Machine Learning . [S.l.] : PMLR , 2021 : 10347 - 10357 . DOI: 10.1109/iccv48922.2021.00091 http://dx.doi.org/10.1109/iccv48922.2021.00091

Li Meng . Transformer-based self-supervised learning and distillation for medical image classification: Improving colorectal cancer detection on NCT-CRC-HE-100K with Swin-T V2 [C ] // Proceedings of 2024 3rd International Conference on Cloud Computing, Big Data Application and Software Engineering (CBASE) . Piscataway : IEEE , 2024 : 644 - 648 . DOI: 10.1109/cbase64041.2024.10824558 http://dx.doi.org/10.1109/cbase64041.2024.10824558

Liu Ze , Hu Han , Lin Yutong , et al . Swin transformer V2: Scaling up capacity and resolution [C ] // 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2022 : 12009 - 12019 . DOI: 10.1109/cvpr52688.2022.01170 http://dx.doi.org/10.1109/cvpr52688.2022.01170

Li A C , Prabhudesai M , Duggal S , et al . Your diffusion model is secretly a zero-shot classifier [C ] // Proceedings of 2023 IEEE/CVF International Conference on Computer Vision . Piscataway : IEEE , 2023 : 2206 - 2217 . DOI: 10.1109/iccv51070.2023.00210 http://dx.doi.org/10.1109/iccv51070.2023.00210

Bhattacharya M , Huang J , Sher A F , et al . Immunodiff: A diffusion model for immunotherapy response prediction in lung cancer [PP/OL ] . V1.arVix ( 2025-05-29 )[ 2026-01-21 ] . https://arxiv.org/abs/2505.23675 https://arxiv.org/abs/2505.23675 .

Chen R J , Ding Tong , Lu M Y , et al . Towards a general-purpose foundation model for computational pathology [J ] . Nature Medicine , 2024 , 30 ( 3 ): 850 - 862 . DOI: 10.1038/s41591-024-02857-3 http://dx.doi.org/10.1038/s41591-024-02857-3

Lu M Y , Chen Bowen , Williamson D F K , et al . A visual-language foundation model for computational pathology [J ] . Nature Medicine , 2024 , 30 ( 3 ): 863 - 874 . DOI: 10.1038/s41591-024-02856-4 http://dx.doi.org/10.1038/s41591-024-02856-4

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

A Method for Enhancing the Transferability of Adversarial Examples Based on Multi-Perspective Confidence Fusion

Image Classification Network with Attenuation Disentangling Mechanism

Global Dependency Guided Feature Reconstruction for Image Classification

Generative Image Detection Based on Diffusion Artifact Contrast Learning

Stacked Rotating Alamouti Code for Quad-Antenna Backscatter Tags

Related Author

ZHAO He-yang

CHEN Ru-hao

LUO Yi-kai

WANG Wen-jian

ZHAO Changfei

DENG Xinyang

JIANG Wen

ZHU Jinbiao

Related Institution

School of Electronics and Information, Northwestern Polytechnical University

Aerospace Information Research Institute, Chinese Academy of Sciences

School of Software, Liaoning Technical University

Key Laboratory of Optoelectronic Information Control and Security Technology

School of Computer Science, School of Cyber Science and Engineering, Nanjing University of Information Science and Technology

⁰