A scale-space curvature matching algorithm for the reconstruction of complex proximal humeral fractures.

摘要: The optimal surgical treatment of complex fractures the proximal humerus is controversial. It proven that best results are obtained if an anatomical reduction fragments achieved and, therefore, computer-assisted methods have been proposed for reconstruction fractures. However, commonly accompanied with a relevant displacement algorithms relying on initial position might fail. state-of-the-art algorithm requires acquisition CT scan (healthy) contralateral anatomy as template to address fragments. Pose-invariant fracture line based applied successful reassembling broken vessels in archaeology. Nevertheless, extraction lines and necessary computation their curvature susceptible noise make application previous approaches difficult or even impossible bone close joints, where cortical layer thin. We present novel scale-space representation curvature, permitting calculate correct alignment between solely corresponding regions lines. automatically reconstructed iterative pairwise validation presented method was performed twelve clinical cases, surgically treated after humeral fracture, by cadaver experiments. accuracy our approach compared humerus. All reconstructions cases resulted accurate approximation pre-traumatic anatomy. outperformed current algorithm.