A fast impingement detection algorithm for computer-aided orthopedic surgery.

ABSTRACT

OBJECTIVE: For simulation of computer-aided orthopedic interventions, the detection of impingement between parts of the patient's anatomy and/or implants is often of key importance. The impingement (collision) detection methods used in the existing literature seem to be unsuitable for two reasons. First, a polyhedral approximation of an anatomical model is not appropriate because medical images are quite irregular and are geometrically complex. Second, geometric and temporal coherences are not always available, because only the final results may be of interest. This article describes the development of a fast and accurate impingement detection algorithm for medical applications. MATERIALS AND METHODS: The presented algorithm takes implicit object models from reconstructions of anatomical CT data that represent complicated anatomical structures. To speed up the detection procedure, a lookup table and a linear transform are used so that searching for impingement between any two objects becomes a problem of calculating spatial indices and checking the lookup table. RESULTS: For any given transformation, the algorithm could perform impingement detection of two objects within 0.1 s on a 167 MHz Sun UltraSPARC1 workstation. Experimental results concerning accuracy, reliability and speed are given for a phantom and for a patient's data set. CONCLUSIONS: This algorithm provides a general-purpose impingement detection method in the sense that objects can be of any shape, and it can be extended to any number of objects in the scene.