Mehmet Zubeyir Unlu

A Deformable Finite Element Model for Breast Image Registration: an Application on Mri / Pet Breast Images Mehmet Zubeyir Unlu

Breast cancer is the most common form of cancer among women. X-ray mammography is the primary and highly sensitive but not very specific way to screen and diagnose it, and in some cases additional imaging and biopsies are necessary. It is desirable to have an alternative non-invasive method to follow up equivocal or difficult to interpret X-ray mammograms, or any inconclusive breast examination. Tomographic medical imaging modalities such as CT, MRI, PET and SPECT provide highly specialized and complementary information to screen and diagnose breast cancer. More complete information about the volumes of interest can be provided by registration and fusion of multiple image volumes created by different modalities. Therefore, image-processing techniques and algorithms are needed to estimate geometrical differences between 3D scans of the same region of interest in different or in the same modalities and to register them. In this book, a new image acquisition and processing protocol, and a method, which relies on a finite element method deformable breast model and a set of fiducial skin markers placed on the breast surface for nonrigid 3D breast image registration is presented.