Design and Assessment of Convolutional Neural Network Based Methods for Vitiligo Diagnosis

Abstract

Background: Today's machine-learning based dermatologic research has largely focused on pigmented/non-pigmented lesions concerning skin cancers. However, studies on machine-learning-aided diagnosis of depigmented non-melanocytic lesions, which are more difficult to diagnose by unaided eye, are very few.Objective: We aim to assess the performance of deep learning methods for diagnosing vitiligo by deploying Convolutional Neural Networks (CNNs) and comparing their diagnosis accuracy with that of human raters with different levels of experience.Methods: A Chinese in-house dataset (2,876 images) and a world-wide public dataset (1,341 images) containing vitiligo and other depigmented/hypopigmented lesions were constructed. Three CNN models were trained on close-up images in both datasets. The results by the CNNs were compared with those by 14 human raters from four groups: expert raters (>10 years of experience), intermediate raters (5–10 years), dermatology residents, and general practitioners. F1 score, the area under the receiver operating characteristic curve (AUC), specificity, and sensitivity metrics were used to compare the performance of the CNNs with that of the raters.Results: For the in-house dataset, CNNs achieved a comparable F1 score (mean [standard deviation]) with expert raters (0.8864 [0.005] vs. 0.8933 [0.044]) and outperformed intermediate raters (0.7603 [0.029]), dermatology residents (0.6161 [0.068]) and general practitioners (0.4964 [0.139]). For the public dataset, CNNs achieved a higher F1 score (0.9684 [0.005]) compared to the diagnosis of expert raters (0.9221 [0.031]).Conclusion: Properly designed and trained CNNs are able to diagnose vitiligo without the aid of Wood's lamp images and outperform human raters in an experimental setting.

Optimizing vitiligo diagnosis with ResNet and Swin transformer deep learning models: a study on performance and interpretability

“…Furthermore, to ensure easy replication and validation of the research methodology, we analyze the proposed methodology in comparison with some state-of-the-art (SOTA) methods that have performed well 29 , 30 . Considering that most of the recent advances in the use of dermoscopic images have been aimed at bridging the gap between clinical and dermoscopic images 31 , 32 .…”

Data-Driven Deep Supervision for Skin Lesion Classification

“…For the public ISIC 2016, ISIC 2017, and ISIC 2018 datasets, we achieve comparable scores with the challenge winners. For both the vitiligo datasets [29], we outperform existing methods.…”

“…Benefits of transfer learning are shown in Section 3. For each patient with suspected vitiligo (e.g., pityriasis alba, hypopigmented nevus), clinical photographs of the affected skin areas were taken by medical assistants using a point-and-shoot camera (as described in [29]). Vitiligo (public) [29]: It contains 1341 images (672 training, 268 validation, and 401 test images).…”

“…Automatic classification of skin lesions such as pigmented/non-pigmented lesions concerning skin cancers [6,1,4,26,23,3] and depigmented and hypomigmented lesions concerning vitiligo (vitiligo vs vitiligo-type, e.g., pityriasis alba, nevus depigmentous [29,13]) in RGB images has garnered lots of research attention in recent years. Convolutional neural network (CNN) based models have shown promising results in accurate classification of skin lesions.…”

Development of an Intelligent Vitiligo Detection Classifier

“…By using Inception-V3, the highest level of accuracy was achieved. • Zhang et al [12] in 2021 Employed CNNs for vitiligo detection and compared the diagnostic accuracy of the CNNs with 14 human experts having different levels of experience. Two datasets, primary (Chinese in-house) as well as secondary (Open Source), were used in this study.…”