湖北汽车工业学院学报2025，Vol.39Issue(2)：56-60,5.DOI:10.3969/j.issn.1008-5483.2025.02.011
基于深度卷积神经网络的超材料微带天线拓扑结构性能预测
Performance Prediction of Metamaterial Microstrip Antenna Topology Based on Deep Convolutional Neural Network
摘要
Abstract
Based on the convolutional neural network model,a metamaterial microstrip antenna topolo-gy layout diagram-gain performance database was established,and the performance prediction of meta-material microstrip antenna topology was achieved.In response to the problem of insufficient prediction accuracy in databases,the convolutional neural network structure model was improved by introducing deep convolutional methods and reverse residual structures.The improved database performance was relatively good,with a loss function reduced to 0.008 and a prediction accuracy of 99%on the valida-tion set,which verified the effectiveness of predicting the performance of metamaterial microstrip anten-na topology based on deep convolutional neural networks.关键词
深度学习/卷积神经网络/超材料/拓扑优化/图像处理Key words
deep learning/convolutional neural network/metamaterial/topological optimization/image processing分类
信息技术与安全科学引用本文复制引用
尹家龙,董焱章,王全伟..基于深度卷积神经网络的超材料微带天线拓扑结构性能预测[J].湖北汽车工业学院学报,2025,39(2):56-60,5.基金项目
国家自然科学基金(11502075) （11502075）
湖北省自然科学基金(2022CFB457) （2022CFB457）
汽车零部件技术湖北省协同创新项目(2015XTZX0401) （2015XTZX0401）
湖北汽车工业学院博士科研启动基金(BK201501) （BK201501）
