Department of Cosmetic Laser Surgery, Hospital for Skin Disease and Institute of Dermatology, Peking Union Medical College and Chinese Academy of Medical Sciences (CAMS), Nanjing, 210042, China.
Jiangsu Key Laboratory of Molecular Biology for Skin Diseases and STIs, Institute of Dermatology, Chinese Academy of Medical Sciences and Peking Union Medical College, Nanjing, 210042, China.
BMC Biotechnol. 2022 Oct 10;22(1):28. doi: 10.1186/s12896-022-00755-5.
OBJECTIVE: We aimed to develop a computer-aided detection (CAD) system for accurate identification of benign pigmented skin lesions (PSLs) from images captured using a digital camera or a smart phone. METHODS: We collected a total of 12,836 clinical images which had been classified and location-labeled for training and validating. Four models were developed and validated; you only look once, v4 (YOLOv4), you only look once, v5 (YOLOv5), single shot multibox detector (SSD) and faster region-based convolutional neural networks (Faster R-CNN). The performance of the models was compared with three trained dermatologists, respectively. The accuracy of the best model was further tested and validated using smartphone-captured images. RESULTS: The accuracies of YOLOv4, YOLOv5, SSD and Faster R-CNN were 0.891, 0.929, 0.852 and 0.874, respectively. The precision, sensitivity and specificity of YOLOv5 (the best model) were 0.956, 0.962 and 0.952, respectively. The accuracy of YOLOv5 model for images captured using a smart-phone was 0.905. The CAD based YOLOv5 system can potentially be used in clinical identification of PSLs. CONCLUSION: We developed and validated a CAD system for automatic identification of benign PSLs using digital images. This approach may be used by non-dermatologists to easily diagnose by taking a photo of skin lesion and guide on management of PSLs.
目的：我们旨在开发一种计算机辅助检测（CAD）系统，以便从使用数码相机或智能手机拍摄的图像中准确识别良性色素性皮肤病变（PSL）。
方法：我们共收集了 12836 张临床图像，这些图像已进行分类和位置标记，用于训练和验证。我们开发并验证了 4 种模型；分别是 You Only Look Once，v4（YOLOv4）、You Only Look Once，v5（YOLOv5）、单发多框检测器（SSD）和更快的基于区域的卷积神经网络（Faster R-CNN）。分别将这 4 种模型的性能与 3 名训练有素的皮肤科医生进行比较。使用智能手机拍摄的图像进一步测试和验证最佳模型的准确性。
结果：YOLOv4、YOLOv5、SSD 和 Faster R-CNN 的准确率分别为 0.891、0.929、0.852 和 0.874。YOLOv5（最佳模型）的精确率、敏感度和特异性分别为 0.956、0.962 和 0.952。使用智能手机拍摄的图像的 YOLOv5 模型准确率为 0.905。基于 CAD 的 YOLOv5 系统可用于临床 PSL 识别。
结论：我们开发并验证了一种使用数字图像自动识别良性 PSL 的 CAD 系统。该方法可由非皮肤科医生使用，通过拍摄皮肤病变的照片来轻松诊断，并指导 PSL 的管理。
PeerJ Comput Sci. 2023-7-3
Sensors (Basel). 2022-6-22
Dermatol Ther (Heidelb). 2025-6-22
Diagnostics (Basel). 2021-7-31
J Digit Imaging. 2020-10
Nat Med. 2020-5-18
J Ethnopharmacol. 2019-8-21
J Drugs Dermatol. 2018-9-1
Int J Dermatol. 2018-7-24
Zotero 插件