Ula? Ba?c?

Automatic Anatomy Recognition and Registration: Medical Image Processing and Analysis Ula? Ba?c?

This thesis tackles two major challenges in medical image analysis: image segmentation and registration. Specifically, we present a novel anatomy recognition algorithm for assisting the image segmentation task, and a novel image registration framework for accurately aligning images. Whilst the anatomy recognition study considers different clinical applications using MR (magnetic resonance) and CT (computed tomography) images, the image registration study uses MR and histological images. We have also developed a novel and effective method for registering histological images of a mouse brain in order to reconstruct the 3D volume. Experiments show that the choice of the reference slice has a significant impact on the quality of 3D reconstruction. Finally, since image intensity variations are inherent in MR images, and since the lack of a standard image intensity scale in MRI leads to many difficulties in tissue characterisability, image display, and analysis, including image segmentation and registration, we systematically investigated and evaluated the influence of intensity standardisation in automatic anatomy recognition and registration tasks using clinical MR images.