衍射
傅里叶变换
光学
微波食品加热
物理
傅里叶分析
衍射层析成像
断层摄影术
人工神经网络
材料科学
微波成像
傅里叶级数
图像处理
计算机科学
信号处理
杂乱
窗口函数
迭代重建
算法
频域
噪音(视频)
叠加原理
振幅
作者
Léo Monnier,Alexandre Baussard,Cyrille Enderli,Guillaume Reille
出处
期刊:Digest
[Institute of Electrical and Electronics Engineers]
日期:2026-07-01
卷期号:: 972-975
标识
DOI:10.1109/ap-s/usnc-ursi60190.2026.11675145
摘要
In this article, we study the potential of PhysicsInformed Neural Operators (PINO) with Fourier Neural Operators (FNO) architecture for electromagnetic scattering in the context of microwave tomography. We consider the TE polarization and dielectric objects. First, we compare the scattered field in the search domain (area surrounding the object) obtained with the considered model and a commercial electromagnetic solver. Next, we compare the scattered field measured by the receivers of the tomographic system. To do this, we calculate the scattered field at the antenna positions from the scattered field obtained at the object's position by the model and the commercial solver. We also evaluate the generalization capabilities of the model by generating the scattered field for new and unseen objects during the training step. The results demonstrate the potential of this approach in terms of accuracy, generalization, and computational time.
科研通智能强力驱动
Strongly Powered by AbleSci AI