Research on a Vitiligo Intelligent Detection Method Based on Convolutional Neural Network
LUO Wei1, WANG Xuelei2, LIU Jian3
, CHEN Jie1, XU Jin1, ZHU Lin1
1. Department of Dermatology, Air Force Medical Center, PLA, Beijing 100142, China; 2. Beijing Electronic Wise Medical Science
and Technology Co., Ltd., Beijing 100085, China; 3. University of Science and Technology Beijing, Beijing 100083, China
Abstract：Objective Based on convolutional neural network, we propose an intelligent method to solve the diagnosis of vitiligo.
Methods Using clinical vitiligo specialist images from the database of the Air Force Medical Center, three different CNN models
(Resnet50, VGG16, Inception v2) were used for model training based on deep learning method. The method was evaluated by
image recognition accuracy, sensitivity, and specificity. Results Compared with traditional image recognition methods based
on single CNN, the proposed method had higher accuracy, precision and sensitivity, and could significantly improve the image
recognition performance of the algorithm. Based on the CNN theoretical model, an intelligent analysis system was successfully
developed for patients with vitiligo. Conclusion This model achieves an effective classification of vitiligo, which helps doctors
reduce the difficulty of work and enables patients to self-diagnose.
罗卫1，王学磊2，刘健3，陈杰1，徐劲1，朱琳1. 基于卷积神经网络的白癜风智能检测研究[J]. 中国医疗设备, 2019, 34(9): 52-54.
LUO Wei1, WANG Xuelei2, LIU Jian3
, CHEN Jie1, XU Jin1, ZHU Lin1. Research on a Vitiligo Intelligent Detection Method Based on Convolutional Neural Network. China Medical Devices, 2019, 34(9): 52-54.
