计算机与现代化Issue(1)：1-6,16,7.DOI:10.3969/j.issn.1006-2475.2026.01.001
基于频域状态空间模型和注意力增强的胃早癌检测
Early Gastric Cancer Detection Based on Frequency Domain State Space Model and Attention Enhancement
摘要
Abstract
Gastric cancer remains a malignancy with high incidence and mortality rates worldwide.Early intervention in gastric cancer can improve patient survival rates,and gastroscopy is the most effective method for screening early gastric cancer.In re-cent years,the use of computer technology to assist medical image processing has rapidly advanced.To enhance the accuracy of early gastric cancer detection,this paper proposes an early gastric cancer detection network based on a Frequency domain State Space model and Attention Enhancement(FSSAE).This method enhances the frequency domain analysis of critical information in gastroscopic images through the Frequency domain State Space model(FSS),improving the recognition accuracy for early gas-tric cancer.Additionally,an Attention Enhancement Module(AEM)is incorporated to improve the network's ability to focus on cancerous regions.A new loss function is also introduced to further enhance network performance.Extensive experiments are con-ducted on the constructed early gastric cancer dataset using the proposed FSSAE network.Ablation experiments validate the ef-fectiveness of FSS and AEM.Comparative experimental results show that the FSSAE network outperforms baseline models and other networks,with the FSSAE's AP metric improving by 2.81%over the baseline network,increasing the detection accuracy for early-stage gastric cancer.关键词
胃早癌检测/频域分析/状态空间模型/注意力增强Key words
early gastric cancer detection/frequency domain analysis/state space model/attention enhancement分类
信息技术与安全科学引用本文复制引用
于玢,戚杏,吴洪磊..基于频域状态空间模型和注意力增强的胃早癌检测[J].计算机与现代化,2026,(1):1-6,16,7.基金项目
山东省自然科学基金资助项目(ZR2022MH230) （ZR2022MH230）
