A computer-implemented method for combining two 3-dimensional ultrasound images of a patient's body part containing blood vessels into a single 3-dimensional ultrasound image of that body part with a greater field of view. Blood vessels within the body part are used for registration. The two ultrasound images are preprocessed using a vessel enhancement algorithm to derive two corresponding enhanced images. The enhanced images can undergo a registration process based upon minimizing a mean-square-difference metric to derive an enhanced rigid transform, which represents a transform that maps the first enhanced image onto the second enhanced image. The two enhanced images are seamlessly combined via a blending process (via the use of a distance transform based weighting of the voxel values) to derive an output image.