Deep Learning-Based Multiclass Framework for Real-Time Melasma Severity Classification: Clinical Image Analysis and Model Interpretability Evaluation

Jun Zhang, Qian Jiang, Qiang Chen, Bin Hu, Liuqing Chen Department of Dermatology, Wuhan No. 1 hospital, Wuhan, Hubei, People’s Republic of ChinaCorrespondence: Bin Hu, Email binhu88@126.com Liuqing Chen, Email chlq35@126.comBackground: Melasma is a prevalent pigmentary disorder characterized by tre...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhang J, Jiang Q, Chen Q, Hu B, Chen L
Format: Article
Language:English
Published: Dove Medical Press 2025-04-01
Series:Clinical, Cosmetic and Investigational Dermatology
Subjects:
Online Access:https://www.dovepress.com/deep-learning-based-multiclass-framework-for-real-time-melasma-severit-peer-reviewed-fulltext-article-CCID
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850171754488528896
author Zhang J
Jiang Q
Chen Q
Hu B
Chen L
author_facet Zhang J
Jiang Q
Chen Q
Hu B
Chen L
author_sort Zhang J
collection DOAJ
description Jun Zhang, Qian Jiang, Qiang Chen, Bin Hu, Liuqing Chen Department of Dermatology, Wuhan No. 1 hospital, Wuhan, Hubei, People’s Republic of ChinaCorrespondence: Bin Hu, Email binhu88@126.com Liuqing Chen, Email chlq35@126.comBackground: Melasma is a prevalent pigmentary disorder characterized by treatment resistance and high recurrence. Existing assessment methods like the Melasma Area and Severity Index (MASI) are subjective and prone to inter-observer variability.Objective: This study aimed to develop an AI-assisted, real-time melasma severity classification framework based on deep learning and clinical facial images.Methods: A total of 1368 anonymized facial images were collected from clinically diagnosed melasma patients. After image preprocessing and MASI-based labeling, six CNN architectures were trained and evaluated using PyTorch. Model performance was assessed through accuracy, precision, recall, F1-score, AUC, and interpretability via Layer-wise Relevance Propagation (LRP).Results: GoogLeNet achieved the best performance, with an accuracy of 0.755 and an F1-score of 0.756. AUC values across severity levels reached 0.93 (mild), 0.86 (moderate), and 0.94 (severe). LRP analysis confirmed GoogLeNet’s superior feature attribution.Conclusion: This study presents a robust, interpretable deep learning model for melasma severity classification, offering enhanced diagnostic consistency. Future work will integrate multimodal data for more comprehensive assessment.Keywords: melasma, deep learning, convolutional neural networks, MASI, clinical decision support
format Article
id doaj-art-247962ea667e480297691185bbebaef2
institution OA Journals
issn 1178-7015
language English
publishDate 2025-04-01
publisher Dove Medical Press
record_format Article
series Clinical, Cosmetic and Investigational Dermatology
spelling doaj-art-247962ea667e480297691185bbebaef22025-08-20T02:20:13ZengDove Medical PressClinical, Cosmetic and Investigational Dermatology1178-70152025-04-01Volume 1810331044102522Deep Learning-Based Multiclass Framework for Real-Time Melasma Severity Classification: Clinical Image Analysis and Model Interpretability EvaluationZhang JJiang QChen QHu BChen LJun Zhang, Qian Jiang, Qiang Chen, Bin Hu, Liuqing Chen Department of Dermatology, Wuhan No. 1 hospital, Wuhan, Hubei, People’s Republic of ChinaCorrespondence: Bin Hu, Email binhu88@126.com Liuqing Chen, Email chlq35@126.comBackground: Melasma is a prevalent pigmentary disorder characterized by treatment resistance and high recurrence. Existing assessment methods like the Melasma Area and Severity Index (MASI) are subjective and prone to inter-observer variability.Objective: This study aimed to develop an AI-assisted, real-time melasma severity classification framework based on deep learning and clinical facial images.Methods: A total of 1368 anonymized facial images were collected from clinically diagnosed melasma patients. After image preprocessing and MASI-based labeling, six CNN architectures were trained and evaluated using PyTorch. Model performance was assessed through accuracy, precision, recall, F1-score, AUC, and interpretability via Layer-wise Relevance Propagation (LRP).Results: GoogLeNet achieved the best performance, with an accuracy of 0.755 and an F1-score of 0.756. AUC values across severity levels reached 0.93 (mild), 0.86 (moderate), and 0.94 (severe). LRP analysis confirmed GoogLeNet’s superior feature attribution.Conclusion: This study presents a robust, interpretable deep learning model for melasma severity classification, offering enhanced diagnostic consistency. Future work will integrate multimodal data for more comprehensive assessment.Keywords: melasma, deep learning, convolutional neural networks, MASI, clinical decision supporthttps://www.dovepress.com/deep-learning-based-multiclass-framework-for-real-time-melasma-severit-peer-reviewed-fulltext-article-CCIDmelasmadeep learningconvolutional neural networksmasiclinical decision support
spellingShingle Zhang J
Jiang Q
Chen Q
Hu B
Chen L
Deep Learning-Based Multiclass Framework for Real-Time Melasma Severity Classification: Clinical Image Analysis and Model Interpretability Evaluation
Clinical, Cosmetic and Investigational Dermatology
melasma
deep learning
convolutional neural networks
masi
clinical decision support
title Deep Learning-Based Multiclass Framework for Real-Time Melasma Severity Classification: Clinical Image Analysis and Model Interpretability Evaluation
title_full Deep Learning-Based Multiclass Framework for Real-Time Melasma Severity Classification: Clinical Image Analysis and Model Interpretability Evaluation
title_fullStr Deep Learning-Based Multiclass Framework for Real-Time Melasma Severity Classification: Clinical Image Analysis and Model Interpretability Evaluation
title_full_unstemmed Deep Learning-Based Multiclass Framework for Real-Time Melasma Severity Classification: Clinical Image Analysis and Model Interpretability Evaluation
title_short Deep Learning-Based Multiclass Framework for Real-Time Melasma Severity Classification: Clinical Image Analysis and Model Interpretability Evaluation
title_sort deep learning based multiclass framework for real time melasma severity classification clinical image analysis and model interpretability evaluation
topic melasma
deep learning
convolutional neural networks
masi
clinical decision support
url https://www.dovepress.com/deep-learning-based-multiclass-framework-for-real-time-melasma-severit-peer-reviewed-fulltext-article-CCID
work_keys_str_mv AT zhangj deeplearningbasedmulticlassframeworkforrealtimemelasmaseverityclassificationclinicalimageanalysisandmodelinterpretabilityevaluation
AT jiangq deeplearningbasedmulticlassframeworkforrealtimemelasmaseverityclassificationclinicalimageanalysisandmodelinterpretabilityevaluation
AT chenq deeplearningbasedmulticlassframeworkforrealtimemelasmaseverityclassificationclinicalimageanalysisandmodelinterpretabilityevaluation
AT hub deeplearningbasedmulticlassframeworkforrealtimemelasmaseverityclassificationclinicalimageanalysisandmodelinterpretabilityevaluation
AT chenl deeplearningbasedmulticlassframeworkforrealtimemelasmaseverityclassificationclinicalimageanalysisandmodelinterpretabilityevaluation