Kinematic Accuracy of Three Surface Registration Methods in a Three-Dimensional Wrist Bone Study

Abstract

The use of registration techniques to determine motion transformations noninvasively has become more widespread with the increased availability of the necessary software. In this study, three surface registration techniques were used to generate carpal bone kinematic results from a single cadaveric wrist specimen. Surface contours were extracted from specimen computed tomography volume images of the forearm, carpal, and metacarpal bones in four arbitrary positions. Kinematic results from each of three registration techniques were compared with results derived from multiple spherical markers fixed to the specimen. Kinematic accuracy was found to depend on the registration method and bone size and shape. In general, rotation errors of the capitate and scaphoid were less than 0.5 deg for all three techniques. Rotation errors for the other bones were generally less than 2 deg, although error for the trapezoid was greater than 2 deg in one technique. Translation errors of the bones were generally less than 1 mm, although errors of the trapezoid and trapezium were greater than 1 mm for two techniques. Tradeoffs existed in each registration method between image processing time and overall kinematic accuracy. Markerless bone registration (MBR) can provide accurate measurements of carpal kinematics and can be used to study the noninvasive, three-dimensional in vivo kinematics of the wrist and other skeletal joints.

Effects of distal radius malunion on distal radioulnar joint mechanics—an in vivo study

“…kinematic error inherent in our techniques for the radius, which include the segmentation and registration algorithms, have been determined to be less than 0.2 AE 0.38 and 0.2 AE 0.1 mm. 29 Third, our joint space area variable (JSA) is not a direct measure of articular contact area. We are limited in our ability to calculate cartilage contact using our current methodologies because cartilage is poorly imaged with CT.…”

Super-resolution registration using tissue-classified distance fields

“…The contour images were then processed with a 3-D imaging software package (Analyze AVW 2.5). The centroids of the seven spherical markers (one set per specimen component) were used to calculate rigid-body motion by a method of least squares [6].…”

Evaluation of achievable registration accuracy of the femur during minimally invasive total hip replacement

“…THE SPATIAL registration between a patient's organ and its computer model is still a critical step in most computer-aided surgery (CAS) applications (GEHRKE et al, 1999;NEU et al, 2000) as a determinant of achievable surgical accuracy. In orthopaedics, there is a trend towards achieving more accurate prosthesis placement, while minimising invasiveness, for procedures including total hip replacement (THR) (WAHRBURG and KERSCHBAUMER, 2000;YAO et al, 2000;YANIV et al, 2000), which remains one of the most common orthopaedic interventions.…”