Automatic location of vertebrae in digitized videofluoroscopic images of the lumbar spine
Automatic location of vertebrae in digitized videofluoroscopic images of the lumbar spine
复制标题
DOI：
10.1016/s1350-4533(96)00050-1
复制
发表时间：
1997-01-01
影响因子：
2.2
通讯作者：
Allen, R
中科院分区：
文献类型：
--
作者：
Muggleton, JM;Allen, R
背痛是一个普遍存在的问题，尽管人们努力控制它，但它所导致的残疾仍在继续增长。背痛诊断和治疗的一个主要问题是评估机械因素发挥作用的程度。在理解这些力学因素中相当重要的是能够量化人体脊柱在体内的实际运动方式。数字化视频透视是目前唯一的实用方法，可用于研究脊柱运动在体内节段水平。脊柱的低剂量、平面、运动X射线被捕获在录像带上，随后被数字化用于分析。到目前为止，数字化图像中的椎骨被手动识别和标记，作为计算椎间运动学的基础。本文介绍了一种自动识别运动序列中的椎骨的过程。该过程增加了客观性和可重复性，并显著减少了在计算运动学之前定位椎骨所需的手动工作。该技术已被应用于校准模型的图像，结果是有希望的。平面内旋转可以被计算到至少1度的精度。重复分析显示椎间旋转的标准差小于0.5度，平移的标准差小于0.25 mm。(C)1997年爱思唯尔科学有限公司为IPEMB。
Back pain is a widespread problem, and the disability it engenders continues to grow, despite efforts to contain it. A major problem in the diagnosis and management of back pain is the assessment of the degree to which mechanical factors play a part. Of considerable importance in understanding these mechanical factors is being able to quantify how the human spine actually moves in vivo. Digitized videofluoroscopy is currently the only practical method available for studying spinal motion in vivo at the segmental level. Low-dose, planar, motion X-rays of the spine are captured on videotape and subsequently digitized for analysis. Until now, vertebrae in the digitized images were identified and marked manually as a basis for calculating intervertebral kinematics. This paper describes a procedure for automatically identifying the vertebrae in the motion sequences. The process increases objectivity and repeatability, and significantly reduces the manual effort required in locating the vertebrae prior to calculating the kinematics. The technique has been applied to images of a calibrated model and the results are promising. In-plane rotations may be calculated to an accuracy of at least 1 degrees. Repeated analysis reveals standard deviations of less than 0.5 degrees for intervertebral rotations and less than 0.25 mm for translations. (C) 1997 Elsevier Science Ltd for IPEMB.