IEEE Trans Biomed Eng. 2020 May;67(5):1328-1337. doi: 10.1109/TBME.2019.2936125. Epub 2019 Aug 19.
Typical microwave tomographic techniques reconstruct the real part of the permittivity with much greater accuracy as compared to the imaginary part. In this paper, we propose a method to mitigate the imbalance between the reconstructed complex permittivity components and increase the accuracy of the overall image recovery. To do so, the complex permittivity in the imaging domain is expressed as a weighted sum of a few preselected permittivities, close to the range of the expected values. To obtain the permittivity weights, a Gauss-Newton algorithm is employed. Image reconstructions from simulated and experimental data for different biomedical phantoms are presented. Results show that the proposed method leads to excellent reconstruction with balanced real and imaginary parts, across different scenarios.
典型的微波层析成像技术在重建介电常数的实部时比重建虚部具有更高的准确性。在本文中，我们提出了一种方法来减轻重建复介电常数分量之间的不平衡，并提高整体图像恢复的准确性。为此，将成像域中的复介电常数表示为接近预期值范围的少数预选介电常数的加权和。为了获得介电常数权重，使用了高斯牛顿算法。针对不同的生物医学模型，分别对模拟和实验数据进行了图像重建。结果表明，该方法在不同情况下都能得到具有良好的实部和虚部平衡的重建结果。
