基于CNN的自动化软组织标志检测系统在二维面部图像上的临床等效性
Clinical Equivalence of a CNN-Based Automated Soft Tissue Landmark Detection System on 2D Facial Images.
摘要(中文译文)
本研究旨在评估并比较基于卷积神经网络（CNN）的深度学习模型与人工标注在正畸应用中二维（2D）面部图像软组织标志点识别方面的准确性、可靠性和时间效率。采用Revopoint Pop2三维扫描仪获取100名18~25岁参与者（50名女性、50名男性）的三维（3D）面部扫描数据，并从3D模型中提取正面和侧面2D图像。由一名研究者使用LabelMe软件进行人工标志点识别，在正面图像上标记22个标志点，在侧面图像上标记15个标志点。基于这些人工标注图像开发并训练了一种新型CNN模型。将该模型的自动标志点识别结果与人工标注在位置误差、识别时间和可重复性方面进行比较。CNN模型的平均定位准确率为96.07%，各解剖点的平均预测误差范围为2.3%~4.5%。发际点（Trichion）、颏前点（Menton）和下颌角点（Gonion）的误差率相对较高。与人工识别相比，该模型显著缩短了标注时间（人工方法：每张图像237秒）。观察者内信度分析显示人工标志点标记具有极好的一致性（ICC：0.85~0.95）。AI模型对相同输入提供了一致的预测结果。基于深度学习的模型在标志点识别准确性方面与人工识别相当，同时显著提高了标注速度和可重复性。这些结果表明，基于CNN的系统为正畸临床分析和数字化工作流程整合提供了一种有前景的替代方案。
Abstract
The aim of this study was to evaluate and compare the accuracy, reliability, and time efficiency of a convolutional neural network (CNN)-based deep learning model with manual annotation in the identification of soft tissue landmarks on two-dimensional (2D) facial images for orthodontic applications. Three-dimensional (3D) facial scans were obtained from 100 participants (50 females, 50 males) aged 18-25 years using the Revopoint Pop2 3D Scanner. Frontal and profile 2D images were extracted from the 3D models. Manual landmark identification was performed by a single investigator using LabelMe software, marking 22 landmarks on frontal images and 15 landmarks on profile images. A novel CNN model was developed and trained on these manually annotated images. The model's automatic landmark identifications were compared with manual annotations in terms of positional error, identification time, and reproducibility. The CNN model achieved a mean localization accuracy of 96.07%. The mean prediction error ranged from 2.3% to 4.5% across various anatomical points. Trichion, Menton, and Gonion points exhibited relatively higher error rates. The model significantly reduced the annotation time compared to manual identification (manual method: 237 s per image). Intra-observer reliability analysis demonstrated excellent agreement for manual landmarking (ICC: 0.85-0.95). The AI model provided consistent predictions for identical inputs. The deep learning-based model demonstrated comparable accuracy to manual landmark identification while significantly improving the annotation speed and reproducibility. These results suggest that CNN-based systems offer a promising alternative for clinical orthodontic analysis and digital workflow integration.
如何引用
AMA
Argun Ege Türkün, Müslim Ege Kalender, Murat Kurt, Servet Doğan. Clinical Equivalence of a CNN-Based Automated Soft Tissue Landmark Detection System on 2D Facial Images.. Diagnostics (Basel, Switzerland). 2026; doi:10.3390/diagnostics16101464.
APA
Argun Ege Türkün, Müslim Ege Kalender, Murat Kurt, Servet Doğan (2026). Clinical Equivalence of a CNN-Based Automated Soft Tissue Landmark Detection System on 2D Facial Images.. Diagnostics (Basel, Switzerland). https://doi.org/10.3390/diagnostics16101464
GB/T 7714
Argun Ege Türkün, Müslim Ege Kalender, Murat Kurt, Servet Doğan. Clinical Equivalence of a CNN-Based Automated Soft Tissue Landmark Detection System on 2D Facial Images.[J]. Diagnostics (Basel, Switzerland), 2026 doi:10.3390/diagnostics16101464.
在 PubMed增强版 阅读全文
翻译全文 · AI 文献问答 · PDF 原文下载 · 引用导出