运动学
人工智能
计算机视觉
计算机科学
软件
运动捕捉
透视
计算机图形学(图像)
生物医学工程
运动(物理)
物理
医学
经典力学
核物理学
程序设计语言
作者
Elizabeth Brainerd,David B. Baier,Stephen M. Gatesy,Tyson L. Hedrick,Keith A. Metzger,Susannah L. Gilbert,Joseph J. Crisco
摘要
X-Ray Reconstruction of Moving Morphology (XROMM) comprises a set of 3D X-ray motion analysis techniques that merge motion data from in vivo X-ray videos with skeletal morphology data from bone scans into precise and accurate animations of 3D bones moving in 3D space. XROMM methods include: (1) manual alignment (registration) of bone models to video sequences, i.e., Scientific Rotoscoping; (2) computer vision-based autoregistration of bone models to biplanar X-ray videos; and (3) marker-based registration of bone models to biplanar X-ray videos. Here, we describe a novel set of X-ray hardware, software, and workflows for marker-based XROMM. Refurbished C-arm fluoroscopes retrofitted with high-speed video cameras offer a relatively inexpensive X-ray hardware solution for comparative biomechanics research. Precision for our biplanar C-arm hardware and analysis software, measured as the standard deviation of pairwise distances between 1 mm tantalum markers embedded in rigid objects, was found to be +/-0.046 mm under optimal conditions and +/-0.084 mm under actual in vivo recording conditions. Mean error in measurement of a known distance between two beads was within the 0.01 mm fabrication tolerance of the test object, and mean absolute error was 0.037 mm. Animating 3D bone models from sets of marker positions (XROMM animation) makes it possible to study skeletal kinematics in the context of detailed bone morphology. The biplanar fluoroscopy hardware and computational methods described here should make XROMM an accessible and useful addition to the available technologies for studying the form, function, and evolution of vertebrate animals.
科研通智能强力驱动
Strongly Powered by AbleSci AI