Key Laboratory of Data Analytics and Optimization for Smart Industry, Northeastern University, Shenyang 110819, China.
Key Laboratory of Infrared Optoelectric Materials and Micro-Nano Devices, Shenyang 110819, China.
Sensors (Basel). 2021 Jun 3;21(11):3869. doi: 10.3390/s21113869.
Image reconstruction of Magnetic induction tomography (MIT) is an ill-posed problem. The non-linear characteristics lead many difficulties to its solution. In this paper, a method based on a Generative Adversarial Network (GAN) is presented to tackle these barriers. Firstly, the principle of MIT is analyzed. Then the process for finding the global optimum of conductivity distribution is described as a training process, and the GAN model is proposed. Finally, the image was reconstructed by a part of the model (the generator). All datasets are obtained from an eight-channel MIT model by COMSOL Multiphysics software. The voltage measurement samples are used as input to the trained network, and its output is an estimate for image reconstruction of the internal conductivity distribution. The results based on the proposed model and the traditional algorithms were compared, which have shown that average root mean squared error of reconstruction results obtained by the proposed method is 0.090, and the average correlation coefficient with original images is 0.940, better than corresponding indicators of BPNN and Tikhonov regularization algorithms. Accordingly, the GAN algorithm was able to fit the non-linear relationship between input and output, and visual images also show that it solved the usual problems of artifact in traditional algorithm and hot pixels in L2 regularization, which is of great significance for other ill-posed or non-linear problems.
磁感应断层成像（MIT）的图像重建是一个不适定问题。其非线性特征给求解带来了许多困难。本文提出了一种基于生成对抗网络（GAN）的方法来解决这些障碍。首先，分析了 MIT 的原理。然后，将寻找电导率分布全局最优值的过程描述为一个训练过程，并提出了 GAN 模型。最后，通过模型的一部分（生成器）对图像进行重建。所有数据集均由 COMSOL Multiphysics 软件的八通道 MIT 模型获得。电压测量样本作为输入提供给训练有素的网络，其输出是对内部电导率分布图像重建的估计。基于所提出的模型和传统算法的结果进行了比较，结果表明，所提出的方法的重建结果的平均均方根误差为 0.090，与原始图像的平均相关系数为 0.940，优于 BPNN 和 Tikhonov 正则化算法的相应指标。因此，GAN 算法能够拟合输入和输出之间的非线性关系，并且可视化图像也表明它解决了传统算法中伪影和 L2 正则化中热点像素的常见问题，这对于其他不适定或非线性问题具有重要意义。
