Image Classification Algorithm for Laryngeal Disease Based on a Multi-Scale Cross-Axis Attention Fusion Network

HU Yansu; WANG Yuanzheng; HOU Jin

doi:10.7652/xjtuxb202606022

您当前的位置：

首页 >

文章列表页 >

Image Classification Algorithm for Laryngeal Disease Based on a Multi-Scale Cross-Axis Attention Fusion Network

更新时间：2026-06-12

- Image Classification Algorithm for Laryngeal Disease Based on a Multi-Scale Cross-Axis Attention Fusion Network
- Journal of Xi'an Jiaotong University Vol. 60, Issue 6, Pages: 256-266(2026)
- 作者机构：
  
  长安大学电子与控制工程学院，710064，西安
  西安交通大学第二附属医院，710114，西安
  陕西省耳鼻咽喉疾病精准诊疗重点实验室，710114，西安
- 作者简介：
- 基金信息：
- DOI：10.7652/xjtuxb202606022
  CLC： TP391.4
- Received：24 October 2025，
  
  Published：10 June 2026
- 稿件说明：
移动端阅览
HU Yansu, WANG Yuanzheng, HOU Jin. Image Classification Algorithm for Laryngeal Disease Based on a Multi-Scale Cross-Axis Attention Fusion Network[J]. Journal of Xi'an Jiaotong University, 2026, 60(6): 256-266.
DOI：

HU Yansu, WANG Yuanzheng, HOU Jin. Image Classification Algorithm for Laryngeal Disease Based on a Multi-Scale Cross-Axis Attention Fusion Network[J]. Journal of Xi'an Jiaotong University, 2026, 60(6): 256-266. DOI： 10.7652/xjtuxb202606022.

摘要

针对咽喉疾病计算机辅助诊断中存在的病灶视角单一、分类精度不足及依赖主观经验判断等问题，提出基于多尺度交叉轴注意力融合网络的咽喉疾病影像算法。引入多尺度交叉轴注意力机制，通过跨空间维度的自适应特征加权，实现细粒度病理特征的精准定位。构建双流协同学习框架，捕获全局解剖特征并聚焦局部上下文信息，实现多视角特征提取。设计特征耦合模块避免简单拼接导致的信息冲突与语义错位，实现特征空间的优化、局部-全局表征的深度整合。研究结果表明：相比于现有的MVT-OFML模型，所提算法在自制单视角、低对比度咽喉影像数据集上的准确性、精确度、召回率、F1分数、特异度可分别获取1.76%、1.56%、3.34%、2.60%、1.10%的性能提升。同时，特征可视化分析揭示了融合网络的决策机制具备显著可解释性，关键病理区域激活强度分布与临床诊断标准保持高度一致。该研究可为智能化辅助诊断系统的开发提供技术支撑。

Abstract

To address limitations in computer-aided diagnosis of laryngeal disease，such as singleview lesion presentation，insufficient classification accuracy，and reliance on subjective experience-based judgment，an image classification algorithm for laryngeal disease based on a multi-scale cross-axis attention fusion network was proposed. A multi-scale cross-axis attention mechanism was introduced to enable precise localization of fine-grained pathological features through adaptive feature weighting across spatial dimensions.A dual-stream collaborative learning architecture was constructed to capture global anatomical structures while focusing on local contextual information，facilitating multi-view feature extraction.A feature coupling module was designed to prevent information conflict and semantic misalignment caused by simple concatenation，optimizing the feature space and deeply integrating local and global representations.Study results showed that，on an in-house single-view，low-contrast laryngeal image dataset，compared with the existing MVT-OFM Lmodel，improvements in accuracy，precision，recall，F1 score，and specificity of 1.76%，1.56%，3.34%，2.60%，and 1.10%，respectively，were achieved.Also，feature visualization analysis revealed that the fusion network’s decision mechanism exhibited noticeable interpretability，with the distribution of activation intensities in key pathological regions being highly consistent with clinical diagnostic criteria.This study is expected to provide technical support for the development of intelligent computer-aided diagnosis systems.

关键词

Keywords

references

付嘉，李丽娟，闫燕，等.深度学习辅助电子喉镜诊断喉白斑的应用研究[J].临床耳鼻咽喉头颈外科杂志，2021，35（5）：464-467.

FU Jia，LI Lijuan，YAN Yan，et al.Application of deep learning assisted electronic laryngoscope in diagnosis of laryngeal leukoplakia[J].Journal of Clinical Otorhinolaryngology Head and Neck Surgery，2021，35（5）：464-467.

马梓博，米悦，张波，等.面向异质性医学图像处理的深度学习算法综述[J].软件学报，2023，34（10）：4870-4915.

MA Zibo，MI Yue，ZHANG Bo，et al.Review on deep learning algorithms for heterogeneous medical image processing[J].Journal of Software，2023，34（10）：4870-4915.

LI Yun，GU Wenxin，YUE Huijun，et al.Real-time detection of laryngopharyngeal cancer using an artificial intelligence-assisted system with multimodal data[J]. Journal of Translational Medicine，2023，21：698.

MA Haomiao，WEI Wenfeng，ZHANG Junhong，et al.Multi-channel deep learning radiomics model based on contrast-enhanced CT for predicting postoperative prognosis in laryngeal carcinoma [J].BMC Cancer，2025，25（1）：1597.

YANG K O，KIM S Y，KANG C W，et al.Diagnosis of unilateral vocal fold paralysis using auto-diagnostic deep learning model [J].Scientific Reports，2025，15（1）：27635.

WANG Haiyang，MAINARDIL.CTGANinaug-mentation of radiomics features classification from narrow band imaging for laryngeal cancer[C]//2024 IEEE International Symposium on Medical Measurements and Applications（MeMeA）.Piscataway，NJ，USA：IEEE，2024：1-5.

VETHASRI V，DEVI AKALYA B，SHRIRAM P M，et al.Deep learning-based classification of laryngeal abnormalities using DenseNet architectures[C]//2025 3rd International Conference on Intelligent Data Communication Technologies and Internet of Things（IDCIoT）. Piscataway，NJ，USA：IEEE，2025：1235-1239.

PATIL S，KIRANGE D.Ensemble of deep learning models for brain tumor detection[J].Procedia Computer Science，2023，218：2468-2479.

SINGH R，GUPTA S，BHARANY S，et al.Ensemble deep learning models for enhanced brain tumor classification by leveraging ResNet50 and EfficientNet-B7 on high-resolution MRI images[J].IEEE Access，2024，12：178623-178641.

THAKUR S，SHARMA S.Ensemble fusion：skin cancer detection using ResNet，EfficientNet，and VGG architectures[C]//2024 15th International Conference on Computing Communication and Networking Technologies（ICCCNT）.Piscataway，NJ，USA：IEEE，2024：1-8.

DUDEJA T，DUBEY S K，BHATT A K.Multinomial classification of CT-MRI image retrieval by optimizing EffResNet-4 architecture in deep neural models[J]. International Journal of System Assurance Engineering and Management，2024，15（8）：3971-3987.

AL-AZZWI Z H N，NAZAROV A N.Brain tumor classification based onimproved stacked ensemble deep learning methods[J].Asian Pacific Journal of Cancer Prevention：APJCP，2023，24（6）：2141-2148.

BHATTACHARJEE R，SUGANYA DEVI K，VIJAYKANTH S.Detecting laryngeal cancer lesions from endoscopy images using deep ensemble model [C]//2023 International Conference on Signal Processing，Computation，Electronics，Power and Telecommunication（IConSCEPT）.Piscataway，NJ，USA：IEEE，2023：1-6.

张龙，胡燕青，赵丽娟，等.采用递归图编码技术与残差网络的滚动轴承故障诊断[J].西安交通大学学报，2023，57（2）：110-120.

ZHANG Long，HU Yanqing，ZHAO Lijuan，et al. Fault diagnosis of rolling bearings using recurrence plot coding technique and residual network[J].Journal of Xi’an Jiaotong University，2023，57（2）：110-120.

GUO M H，LU C Z，HOU Q，et al.Segnext：rethinking convolutional attention design for semantic segmentation[J].Advances in Neural Information Processing Systems，2022，35：1140-1156.

SHAO Hao，ZENG Quansheng，HOU Qibin，et al. MCANet：medical image segmentation with multiscale cross-axis attention [J]. Machine Intelligence Research，2025，22（3）：437-451.

TAN Mingxing，LE W.EfficientNet：rethinking model scaling for convolutional neural networks[C]//Proceedings of the 36th International Conference on Machine Learning.Chia Laguna Resort，Sardinia，Italy：PMLR，2019：6105-6114.

HAO Jie，LIU Yangrui，MO Zhiqiang，et al.Integrated multi-omics and whole slide images for survival prediction in glioblastoma using multiple instance learning and co-attention[C]//2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society（EMBC）.Piscataway，NJ，USA：IEEE，2024：1-4.

赵小明，廖越辉，张石清，等.融合局部与全局特征的DCE-MRI乳腺肿瘤良恶分类[J].中国图象图形学报，2024，29（1）：256-267.

ZHAO Xiaoming，LIAO Yuehui，ZHANG Shiqing，et al. Method of classifying benign and malignant breast tumors by DCE-MRI incorporating local and global features [J].Journal of Image and Graphics，2024，29（1）：256-267.

MEHTA S，RASTEGARI M. Mobil eViT：lightweight，general-purpose，and mobile-friendly vision transformer [EB/OL].（2022-03-04）[2025-09-04]. https：//arxiv.org/abs/2110.02178.

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.

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.

HUANG Gao，LIU Zhuang，VAN DER MAATEN L，et al.Densely connected convolutional networks [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition（CVPR）.Piscataway，NJ，USA：IEEE，2017：2261-2269.

HOWARD A G，ZHU Menglong，CHEN Bo，et al. MobileNets：efficient convolutional neural networks for mobile vision applications[EB/OL].（2017-04-17）[2025-08-24].https：//arxiv.org/abs/1704.04861.

HU Jie，SHEN Li，SUN Gang.Squeeze-and-excitation networks [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway，NJ，USA：IEEE，2018：7132-7141.

CAI Xinhao，LAI Qiuxia，WANG Yuwei，et al.Poly kernel inception network for remote sensing detection [C]//2024 IEEE/CVFConference on Computer Vision and Pattern Recognition（CVPR）.Piscataway，NJ，USA：IEEE，2024：27706-27716.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Bi-Level Multi-Obj ective Optimization of a Wind-Solar Coupled Hydrogen Production System Based on the Synergy of Hybrid Electrolyzer Arrays

Research on the Electrohydrodynamic Effects of Ionic Liquid-Modified Fuels

Study on the Thermodynamic Properties in the Anomalous Region of Supercritical Water

Diffusion Characteristics and Influencing Factors in Case of Hydrogen Leakage from Fuel Cell Vehicles Running in Tunnels

Numerical Simulation of Heat-Moisture Barrier Performance of a Horizontal Air Curtain in Cold Storage

Related Author

HU Yansu

WANG Yuanzheng

HOU Jin

Related Institution

School of Electronic and Control Engineering, Chang’an University

Department of Otorhinolaryngology-Head and Neck Surgery, The Second Affiliated Hospital of Xi’an Jiaotong University

⁰