Large-Deformation Image Registration of CT-TEE for Surgical Navigation of Congenital Heart Disease

Algorithm 1

Procedure of the proposed method.

Input: Cardiac CT image and TEE image

Output:Result of final registration

Procedure of the Proposed Method:

Step 1: Cardiac CT and TEE images were segmented interactively to yield the CT and TEE

image segmentations separately.

Step 2: During interactive segmentations, the spatial positions of heart valves were used as priori information.

Step 3: Based on the valve data, CT and TEE images were registered to yield the transformation

matrix and the results of the basic registration .

Step 4: ROI enhancement was performed for the and to obtain the enhanced CT and the

enhanced TEE images. During segmentation, gray-scale enhancement was performed to obtain the

enhanced segmentations of CT and TEE images.

Step 5: The probability map of the CT image was generated based on the enhanced CT image

and its enhanced segmentation . The probability map of TEE image was generated similarly.

Step 6: Finally, the probability maps and were used to perform the final registration by maximizing

similarity, producing the final registration transformation and the result of final registration .