Abstract
Objective: The Achilles tendon (AT) is the longest and strongest tendon in the human body. Ultrasound (US) imaging including B-mode US and sonoelastography is valuable for examination of AT. Due to its long length, the AT is often divided into multiple successive and overlapping sections for US scanning. We aim to propose a coarse-to-fine method for image stitching of these multiple US scans.
Methods: First, a new similarity measure for rigid registration is proposed based on the boundary vector field, and it is maximized by using the particle swarm optimization for coarse alignment. Second, the log-domain diffeomorphic registration is followed for non-rigid refinement. Finally, the weighted averaging is utilized for image blending.
Results: Simulation experiments of non-rigid registration on 20 image pairs show that the horizontal, vertical and distance registration errors of our method are 1.422, 0.628, and 1.616 pixel, respectively. Non-rigid registration experiments on 11 pairs of real clinical US images demonstrate that the horizontal, vertical and distance registration errors are 1.514, 1.205, and 1.928 pixel, respectively. Registration results on both synthetic images and real clinical images by our method exceeded those of other four methods. After image blending on multiple AT sections of US scans, panoramas of AT are generated for dual modality visualization of both B-mode and sonoelastography.
Conclusions: Our technique can accurately stitch multiple sections of both B-mode and elastographic US scans and provide panoramas of AT, which may help diagnosis and treatment evaluation of tendinopathies and tendon ruptures.
Published on: December 19, 2017
doi: 10.17756/micr.2017-003
Citation: Xiong J, Zhang Q, Suo J, Cai Y, Shi J, et al. 2017. Image Stitching for Achilles Tendon Ultrasonography Using Particle Swarm Optimization of Boundary Vector Field Based Similarity and Log-Domain Diffeomorphic Registration. J Med Imaging Case Rep 1(1): 8-15.
