Experimental Validation of an Elastic Registration Algorithm for Ultrasound Images

Ultrasound is a favorable tool for intra-operative surgical guidance due to its fastimaging speed and non-invasive nature. However, deformations of the anatomy causedby breathing, heartbeat, and movement of the patient make it difficult to track thelocation of anatomical landmarks during intra-operative ultrasound-guided interven-tions. While elastic registration can be used to compensate for image misalignment,its adaptation for clinical use has only been gradual due to the lack of standardizedguidelines to quantify the performance of different registration techniques.Evaluation of elastic registration algorithms is a difficult task since the pointto point correspondence between images is usually unknown. This poses a majorchallenge in the validation of non-rigid registration techniques for performance com-parisons. Current validation guidelines for non-rigid registration algorithms existfor the comparison of techniques for magnetic resonance images of the brain. Theseframeworks provide users with standardized brain datasets and performance measuresbased on brain region alignment, intensity differences between images, and inverseconsistency of transformations. These metrics may not all be suitable for ultrasoundregistration algorithms due to the different properties of the imaging modalities. Fur-thermore, other metrics are required for validating the registration performance ondifferent anatomical images with large deformations such as the liver.i This work presents a validation framework dedicated for ultrasound elastic reg-istration algorithms. Quantitative validation metrics are evaluated for ultrasoundimages. These include a simulation technique to measure registration accuracy, asegmentation algorithm to extract anatomical landmarks to measure feature overlap,and a technique to measure the alignment of images using similarity metrics.An extensive study of an ultrasound temporal registration algorithm is conductedusing the proposed validation framework. Experiments are performed on a largedatabase of 2D and 3D US images of the carotid artery and the liver to assess theperformance of this algorithm. In addition, two graphical user interfaces which in-tegrate the image registration and segmentation techniques have been developed tovisualize the performance of these algorithms on ultrasound images captured in realtime. In the future, these interfaces may be used to enhance ultrasound examination.