Multi-stage Registration for Quantification of Lung Perfusion in Chest CT Images
We propose a multi-stage registration technique to identify perfusion defects of the lungs using pre- and post-contrast CT images. Our method is composed of four main steps. First, point-based rough registration is performed for correcting gross translational mismatch. The center of inertia (COI), apex and hilar point of each unilateral lung are proposed as the reference point. Second, the initial alignment is refined by narrow-band surface registration. Third, thin-plate spline warping is used to accurately align inner region of the lung. Finally, enhanced vessels are visualized by subtracting registered pre-contrast images from post-contrast images. To facilitate visualization of parenchymal enhancement, color-coded mapping and image fusion is applied to subtracted images. Our method has been evaluated using eight pairs of pre- and post-contrast CT images by visual inspection, accuracy evaluation and processing time.
KeywordsLung Perfusion Parenchymal Enhancement Surface Registration Compute Tomography Perfusion Image Chest Compute Tomography Image
Unable to display preview. Download preview PDF.
- 4.Zhou, C., Hadjiisk, L.M., Sahiner, B., et al.: Computerized detection of pulmonary embolism in 3D computed tomographic images: vessel tracking and segmentation technique. In: Proc. of SPIE Medical Imaging, vol. 5032, pp. 1613–1620 (2003)Google Scholar
- 8.Kim, N.K., Seo, J.M., Heo, J.N., Hong, H., Lee, J.J., Kang, S.H.: Analysis of point-to-point lung motion with full inspiration and expiration CT data: optimal geometric assumption model for the effective registration algorithm. In: Proc. of RSNA 2005 (2005)Google Scholar