论文部分内容阅读
目的:探讨计算机辅助颈椎分析系统(CACVA,Computer Aided Cervical Vertebra Analysis)半自动识别颈椎标志点的准确性及重复性。方法:随机选取口腔正畸门诊青少年患者常规拍摄的X线头颅侧位定位片120张,由专家组、正畸初学者组、计算机辅助颈椎分析组定点第2～5颈椎体的27个标志点,各组分别定点三次,比较三组之间定点的准确性和组内重复性。专家组和初学者组采用手工定点,CACVA组采用快速行进算法和抛物线拟合来分割颈椎图像,通过多段曲线拟合的方法对颈椎骨边缘拟合,进而计算出标记点。随机区组方差分析及两两比较进行三组定点准确性和重复性的比较。结果:从定点准确性来看,CACVA组与专家组差异无显著性(P>0.05),计算机辅助颈椎分析组与初学组、专家组和初学组差异均有显著性(P<0.05);从定点重复性来看,CAVCA组优于专家组和初学组。结论:CAVCA组与专家组手工定点具有同样的准确性,同时具有重复性好,简便快速的优势,值得正畸临床和科研推广。
Objective: To investigate the accuracy and repeatability of semi-automatic identification of cervical vertebra mark by Computer Aided Cervical Vertebra Analysis (CACVA). Methods: Totally 120 radiographs of X-ray cephalometric images were randomly selected from outpatients with orthodontic treatment. The expert group, orthodontics group and computer-aided cervical spine analysis group were given 27 points , Each group were fixed point three times, comparing the accuracy of the three groups of fixed-point and intra-group repeatability. The expert group and the beginner group were manually fixed. The CACVA group was divided into two groups by fast-moving algorithm and parabola fitting. The cervical vertebra images were fitted by multiple curve fitting method, and then the marked points were calculated. Randomized block analysis of variance and pairwise comparison of three groups of fixed-point accuracy and repeatability comparison. Results: There was no significant difference between the CACVA group and the expert group (P> 0.05). The differences between the computer-assisted cervical spine analysis group and the beginner group, the expert group and the beginner group were significant (P <0.05) Fixed-point repeatability point of view, CAVCA group is better than the expert group and beginner group. Conclusion: The CAVCA group has the same accuracy as the manual grouping by the expert group, meanwhile it has the advantages of good repeatability, simple and quickness, which is worth to be popularized clinically and scientifically.