Deep Learning for the Radiographic Detection of Periodontal Bone Loss
Deep Learning for the Radiographic Detection of Periodontal Bone Loss
复制标题
DOI：
10.1038/s41598-019-44839-3
复制
发表时间：
2019-06-11
影响因子：
4.6
通讯作者：
Schwendicke, Falk
中科院分区：
文献类型：
--
作者：
Krois, Joachim;Ekert, Thomas;Schwendicke, Falk
我们应用深度卷积神经网络（CNN）来检测全景牙科X光片上的牙周骨丢失（PBL）。我们合成了一组2001年的图像片段从全景射线照片。我们的参考测试是PBL的测量%。深度前馈CNN通过10次重复的组洗牌进行训练和验证。使用网格搜索调整模型架构和超参数。最终的模型是一个七层深度神经网络，由4，299，651个权重参数化。为了比较，六位牙医评估PBL的图像片段。平均超过10个验证倍数，CNN的平均（SD）分类准确度为0.81（0.02）。平均（SD）敏感性和特异性分别为0.81（0.04），0.81（0.05）。牙医的平均（SD）准确度为0.76（0.06），但CNN与检查员相比无统计学显著性上级（p = 0.067/t检验）。牙医的平均敏感性和特异性分别为0.92（0.02）和0.63（0.14）。在有限数量的射线照相图像片段上训练的CNN显示出至少与牙医在全景射线照相上评估PBL相似的辨别能力。牙医在使用X光片时的诊断工作可以通过应用基于机器学习的技术来减少。
We applied deep convolutional neural networks (CNNs) to detect periodontal bone loss (PBL) on panoramic dental radiographs. We synthesized a set of 2001 image segments from panoramic radiographs. Our reference test was the measured % of PBL. A deep feed-forward CNN was trained and validated via 10-times repeated group shuffling. Model architectures and hyperparameters were tuned using grid search. The final model was a seven-layer deep neural network, parameterized by a total number of 4,299,651 weights. For comparison, six dentists assessed the image segments for PBL. Averaged over 10 validation folds the mean (SD) classification accuracy of the CNN was 0.81 (0.02). Mean (SD) sensitivity and specificity were 0.81 (0.04), 0.81 (0.05), respectively. The mean (SD) accuracy of the dentists was 0.76 (0.06), but the CNN was not statistically significant superior compared to the examiners (p = 0.067/t-test). Mean sensitivity and specificity of the dentists was 0.92 (0.02) and 0.63 (0.14), respectively. A CNN trained on a limited amount of radiographic image segments showed at least similar discrimination ability as dentists for assessing PBL on panoramic radiographs. Dentists' diagnostic efforts when using radiographs may be reduced by applying machine-learning based technologies.