Groupwise registration of dynamic cardiac perfusion images using temporal dynamics and segmentation information

We propose a groupwise registration method for cardiac perfusion images using temporal information. Perfusion image characteristics make groupwise registration especially attractive, and registration accuracy is improved by including segmentation information. Registration aims to maximize the smoothness of the intensity signal while segmentation minimizes a pixel’s intensity difference with other pixels having the same segmentation label. We use B-splines to optimize the cost function that includes registration and segmentation information. The method was tested on real patient datasets. Due to the use of temporal information for groupwise registration our method shows lower registration error and higher segmentation accuracy than conventional methods which perform registration and segmentation separately.