提高时域乳腺光学层析成像质量的方法

Enhancing Imaging Quality of Time-Domain Optical Mammography

时域乳腺光学层析成像技术可以有效重建乳腺组织的光学参数,实现乳腺癌的早期检测.为了提高重建图像的性能,针对图像重建过程中吸收系数和约化散射系数的雅克比矩阵间的量级差异,提出了一种有效的雅克比矩阵标定方法;为了克服不适定性因素对重建图像质量的影响,引入了图像分割技术实现对雅克比矩阵有效降维.实验数据的相关验证表明,上述两种方法可有效提高重建图像的质量.

Time-domain optical mammography can effectively reconstruct optical parameters of the breast tissue and achieve early diagnosis of breast cancer.To improve the performance of the reconstructed images,an effective Jacobian matrix scaling method was proposed according to the magnitude difference between Jacobian matrix of absorption coefficient and that of reduced scattering coefficient in image reconstruction.An image segmentation technique was introduced to overcome the influence of the ill-posedness on the quality of reconstructed images and to remarkably reduce the dimension of Jacobian matrix.Experimental results reveal that Jacobian matrix scaling and image segmentation can effectively enhance the quality of the reconstructed images.