Non-rigid registration using free-form deformation for prostate images

Precise image alignment is of practical interest in clinical diagnosis and treatment. In this paper, a non-rigid registration algorithm is developed in order to find the global and local motion of the prostate tissue. Modeling local deformation of prostate tissue, among other clinical applications, can improve the dosimetry planning in prostate brachytherapy. This model is obtained using free-form deformation (FFD) based on B-splines. The similarity measure used in this approach is normalized mutual information. The algorithm is applied to 2D ultrasound images of prostate for fully automated registration. The algorithm led to good results in terms of computing both global and local misregistration in prostate images.