论文部分内容阅读
乳腺癌是危害妇女健康的高死亡率肿瘤疾病,乳腺筛查是当前降低死亡率的有效途径。时域微波成像比钼靶X线等常规技术具有费用低、辐射小等优势,更适合筛查发现早期的乳腺肿瘤。然而,成像问题的非线性、病态特性以及乳腺组织的色散特性是该技术面临的三个主要困难。为此,本文提出了一种时域微波层析成像新技术,分别通过非线性最优化方法、正则化方法、Cole-Cole建模方法,有效克服了这些困难。在一个二维的数值算例中,考虑了噪声影响,仿真结果初步证实:该技术检测小尺寸乳腺肿瘤是可行的;易于发现浅层肿瘤;且反演静态电导率的精度最高。
Breast cancer is a high-risk cancer disease that endangers women’s health. Breast screening is an effective way to reduce mortality. Time-domain microwave imaging than the mammography X-ray and other conventional techniques with low cost, radiation and other advantages, more suitable for screening for early detection of breast cancer. However, the non-linearity of the imaging problem, the pathological features, and the dispersion characteristics of breast tissue are the three major difficulties that the technology faces. To this end, this paper presents a new time-domain microwave tomography technology, respectively, through the nonlinear optimization method, regularization method, Cole-Cole modeling methods, effectively overcome these difficulties. In a two-dimensional numerical example, the influence of noise was considered. The simulation results initially confirmed that this technique is feasible to detect small breast tumors; it is easy to find superficial tumors; and the precision of static conductivity inversion is the highest.