论文部分内容阅读
目的提高早期邻面龋的临床诊断水平,建立邻面龋X线图像的计算机自动分析诊断系统(CDS),并对诊断效果进行初步评价。方法对含有早期邻面龋损、模拟天然牙列的95颗离体牙,进行平行投照的翼片检查并获得X线图像。通过对邻面龋损特征的分析比较,运用Matlab和VC语言,建立CDS。应用该系统自动检测牙齿邻面是否存在龋损及其深度。同时由8位检查者阅片。以组织学结果为金标准,评价诊断的准确度、敏感度、特异度及一致性等水平。结果CDS诊断的敏感度为0·728,阅片法的平均敏感度为0·543。系统诊断深度与组织学深度的Pearson相关系数为0·722,优于肉眼阅片与组织学深度的相关系数(0·584)。系统重复检测的一致性较好。结论CDS可以对邻面早期龋损进行自动诊断,其总体诊断水平优于肉眼阅片。
Objective To improve the clinical diagnosis of early carious proximal caries and to establish a computerized automatic diagnosis system (CDS) of caries on the adjacent caries and to evaluate the diagnostic value. Methods Ninety-five isolated teeth with early adjacent carious lesions and simulated natural dentition were examined by parallel projection and X-ray images were obtained. Through the analysis of the characteristics of the adjacent carious lesions, the use of Matlab and VC language, the establishment of CDS. The system automatically detects the presence of carious lesions and their adjacent teeth. At the same time review by 8 inspectors. To histological results as the gold standard, evaluation of diagnostic accuracy, sensitivity, specificity and consistency and other levels. Results The sensitivity of CDS diagnosis was 0.728, and the average sensitivity of reading method was 0.543. The Pearson correlation coefficient between the depth of systematic diagnosis and the depth of histology was 0.722, which was better than the correlation coefficient of the depth of the naked eye and the depth of histology (0. 584). The consistency of the system repeated testing is better. Conclusion CDS can diagnose the early caries of the adjacent face automatically, and its overall diagnosis is better than that of the naked eye.