Model-based tracking of the bones of the foot: A biplane fluoroscopy validation study. (1st January 2018)

• Record Type: Journal Article
• Authors: Iaquinto, Joseph M.; Kindig, Matthew W.; Haynor, David R.; Vu, QuocBao; Pepin, Nathan; Tsai, Richard; Sangeorzan, Bruce J.; Ledoux, William R.
• Is Part Of: Computers in biology and medicine. Volume 92(2018)
• Page Start: 118
• Page End: 127
• Publication Date: 2018-01-01
• Subjects: Biplane fluoroscopy -- Model-based -- Bone-based -- Validation -- Foot and ankle; Medicine -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; 610.285
• DOI: 10.1016/j.compbiomed.2017.11.006
• Languages: English
• ISSNs: 0010-4825

Abstract

Abstract: Measuring foot kinematics using optical motion capture is technically challenging due to the depth of the talus, small bone size, and soft tissue artifact. We present a validation of our biplane X-ray system, demonstrating its accuracy in tracking the foot bones directly. Using an experimental linear/rotary stage we imaged pairs of tali, calcanei, and first metatarsals, with embedded beads, through 30 poses. Model- and bead-based algorithms were employed for semi-automatic tracking. Translational and rotational poses were compared to the experimental stage (a reference standard) to determine registration performance. For each bone, 10 frames per pose were analyzed. Model-based: The resulting overall translational bias of the six bones was 0.058 mm with a precision of ± 0.049 mm. The overall rotational bias of the six bones was 0.42° with a precision of ± 0.41°. Bead-based: the overall translational bias was 0.037 mm with a precision of ± 0.032 mm and for rotation was 0.29° with a precision of ± 0.26°. We validated the accuracy of our system to determine the spatial position and orientation of isolated foot bones, including the talus, calcaneus, and first metatarsal over a range of quasi-static poses. Although the accuracy of dynamic motion was not assessed, use of an experimental stage establishes a reference standard. Highlights: A model-based bone tracking algorithm is validated for foot and ankle bones. The translational accuracy of the system was 0.058 mm with a precision of ±0.049 mm. The rotational accuracy of the system was 0.42° with a precision of ±0.41°. Comparable bead-tracking errors support the use of precision stages in validation.