摘要
鉴于乳腺光学层析成像方法中的不适定性严重影响重建图像性能,引入图像分割技术.通过选择合适的图像分割阈值,将基于全域重建的图像分割为背景区域和目标区域,再对得到的"目标区域"进行二次图像重建.数值模拟研究表明,该技术可有效地减小反演变量的维数,提高重建图像的性能.
Considering that the reconstructed image is seriously impacted by the ill-posedness of optical mammography,image segmentation technique is introduced. Choosing an appropriate threshold of image segmentation, the image reconstructed on the overall region can be divided into the background region and the target region. The second image reconstruction is limited in the target region. The simulative experiments reveal that the performance of reconstructed image can be improved by efficiently reduced the dimension of the parameters to be reconstructed.
出处
《光子学报》
EI
CAS
CSCD
北大核心
2009年第9期2423-2429,共7页
Acta Photonica Sinica
基金
国家重点研究发展计划(2006CB705700)
国家自然科学基金(30870657
60678049)
天津市自然科学基金(07JCYBJC06600)
高等学校学科创新引智计划(B07014)资助
关键词
乳腺癌早期检测
时域扩散光学成像
有限元法
图像分割技术
Breast cancer earlier examination
Time-domain diffuse optical tomography
Finite-element method
Image segmentation technique