皮肤病深度学习诊断模型的研究进展

doi:10.13481/j.1671-587X.20250633

吉林大学学报(医学版) ›› 2025, Vol. 51 ›› Issue (6): 1755-1762.doi: 10.13481/j.1671-587X.20250633

• 综述 • 上一篇

皮肤病深度学习诊断模型的研究进展

丛玉佳¹,刘冰²,苗百卉¹,丛宪玲³(),姜日花¹()

^1.吉林大学中日联谊医院皮肤性病科，吉林长春 130033
^2.吉林大学中日联谊医院病理科，吉林长春 130033
^3.吉林大学中日联谊医院生物样本库，吉林长春 130033

收稿日期:2025-01-13 接受日期:2025-02-21 出版日期:2025-11-28 发布日期:2025-12-15
通讯作者: 丛宪玲,姜日花 E-mail:congxl@jlu.edu.cn;jiangrh@jlu.edu.cn
作者简介:丛玉佳（2000-），女，吉林省榆树市人，在读硕士研究生，主要从事皮肤性病学疾病与人工智能诊断交叉融合方面的研究。
基金资助:
吉林省科技厅技术创新引导国际科技合作项目(20240402002GH);吉林省卫生健康科技能力提升计划项目(202WS-YA010)

Research progress in deep learning-based diagnostic models for dermatological diseases

Yujia CONG¹,Bing LIU²,Baihui MIAO¹,Xianling CONG³(),Rihua JIANG¹()

^1.Department of Dermatology，China-Japan Union Hospital，Jilin University，Changchun 130033，China
^2.Department of Pathology，China-Japan Union Hospital，Jilin University，Changchun 130033，China
^3.Department of Biobank，China-Japan Union Hospital，Jilin University，Changchun 130033，China

Received:2025-01-13 Accepted:2025-02-21 Online:2025-11-28 Published:2025-12-15
Contact: Xianling CONG,Rihua JIANG E-mail:congxl@jlu.edu.cn;jiangrh@jlu.edu.cn

摘要/Abstract

摘要：

皮肤病严重影响全球约1.9亿患者的生活质量，其疾病表现的复杂性和多样性是传统诊疗模式面临的挑战，探索新型的诊断手段已经成为当前亟待解决的关键任务。深度学习（DL）技术在皮肤病智能识别领域已逐渐广泛应用，并显示出巨大的潜力。本文从三大维度系统综述了DL在皮肤病诊断中的研究进展。首先，在数据输入层面，重点分析皮肤镜图像、超声影像及病理切片图像多模态数据的特征提取与预处理方法。其次，在算法模型层面，深入探讨集成学习框架、多模态数据融合策略、多中心数据集协同训练方法及可解释性模型构建方案。最后，在任务识别层面，评估DL模型在良性皮肤病筛查、恶性皮肤病鉴别以及二分类与多分类诊断任务中的性能表现。通过多角度评述皮肤病DL诊断模型的研究为未来模型的进一步优化与应用提供参考。

关键词: 深度学习, 皮肤病, 数据输入, 智能识别, 皮肤镜图像, 超声影像

Abstract:

Skin diseases significantly affect the quality of life of approximately 190 million individuals worldwide. The complexity and diversity of their clinical manifestations are the major challenges for traditional diagnostic approaches， and exploring novel diagnostic strategies has become an urgent priority. In recent years， deep learning （DL） technology has been increasingly applied in the intelligent recognition of skin diseases， demonstrating substantial potential. This study provides a systematic review of the research progress of DL in dermatological diagnosis from three major dimensions. First， at the data input level， it focuses on the characterization and preprocessing of multimodal data， including dermoscopic images， ultrasound images， and histopathological slides. Second， at the algorithmic model level， it explores ensemble learning frameworks， multimodal data fusion strategies， multicenter collaborative training approaches， and interpretable model construction. Finally， at the task recognition level， it evaluates the performance of DL models in benign skin disease screening， malignant skin lesion differentiation， and binary as well as multiclass classification tasks. By comprehensively reviewing advancements in DL-based skin disease diagnostic models from multiple perspectives， this paper aims to provide valuable insights for the further optimization and clinical translation of intelligent diagnostic systems.

Key words: Deep learning, Skin diseases, Input data, Intelligent identification, Dermoscopic images, Ultrasound images

中图分类号:

R751

丛玉佳,刘冰,苗百卉,丛宪玲,姜日花. 皮肤病深度学习诊断模型的研究进展[J]. 吉林大学学报(医学版), 2025, 51(6): 1755-1762.

Yujia CONG,Bing LIU,Baihui MIAO,Xianling CONG,Rihua JIANG. Research progress in deep learning-based diagnostic models for dermatological diseases[J]. Journal of Jilin University(Medicine Edition), 2025, 51(6): 1755-1762.

参考文献 53

[1]	POALELUNGI D G， MUSAT C L， FULGA A， et al. Advancing patient care： how artificial intelligence is transforming healthcare ［J］. J Pers Med， 13（8）： 1214.
[2]	LI M， LIU C， PAN X， et al. Digital twin-assisted graph matching multi-task object detection method in complex traffic scenarios ［J］. Sci Rep， 2025， 15（1）： 10847.
[3]	FANG M， WANG Z， PAN S， et al. Large models in medical imaging： advances and prospects ［J］. Chin Med J （Engl）， 2025， 138（14）： 1647-1664.
[4]	GUO T Y， DENG J H， ZHOU Z M， et al. Advancements in deep learning-based image screening for orthopedic conditions： emphasis on osteoporosis， osteoarthritis， and bone tumors ［J］. Ageing Res Rev， 2025， 111： 102840.
[5]	ZHU X， ZHENG B， CAI W， et al. Deep learning-based diagnosis models for onychomycosis in dermoscopy ［J］. Mycoses， 2022， 65（4）： 466-472.
[6]	YU Z， KAIZHI S， JIANWEN H， et al. A deep learning-based approach toward differentiating scalp psoriasis and seborrheic dermatitis from dermoscopic images［J］. Front Med （Lausanne）， 2022， 9： 965423.
[7]	AEEM A， ANEES T， FIZA M， et al. SCDNet： a deep learning-based framework for the multiclassification of skin cancer using dermoscopy images ［J］. Sensors （Basel）， 2022， 22（15）： 5652.
[8]	PHAM T C， LUONG C M， HOANG V D， et al. AI outperformed every dermatologist in dermoscopic melanoma diagnosis， using an optimized deep-CNN architecture with custom mini-batch logic and loss function［J］. Sci Rep， 2021， 11（1）： 17485.
[9]	LEE S， CHU Y S， YOO S K， et al. Augmented decision-making for acral lentiginous melanoma detection using deep convolutional neural networks［J］. J Eur Acad Dermatol Venereol， 2020， 34（8）： 1842-1850.
[10]	LAVERDE-SAAD A， JFRI A， GARCÍA R， et al. Discriminative deep learning based benignity/malignancy diagnosis of dermatologic ultrasound skin lesions with pretrained artificial intelligence architecture［J］. Skin Res Technol， 2022， 28（1）： 35-39.
[11]	CZAJKOWSKA J， BADURA P， KORZEKWA S， et al. Deep learning-based high-frequency ultrasound skin image classification with multicriteria model evaluation［J］. Sensors （Basel）， 2021， 21（17）： 5846.
[12]	CZAJKOWSKA J， JUSZCZYK J， PIEJKO L， et al. High-frequency ultrasound dataset for deep learning-based image quality assessment［J］. Sensors （Basel）， 2022， 22（4）： 1478.
[13]	LEE H， LEE Y， JUNG S W， et al. Deep learning-based evaluation of ultrasound images for benign skin tumors［J］. Sensors （Basel）， 2023， 23（17）： 7374.
[14]	CZAJKOWSKA J， BADURA P， KORZEKWA S， et al. Deep learning approach to skin layers segmentation in inflammatory dermatoses［J］. Ultrasonics， 2021， 114： 106412.
[15]	LEE S， RAHUL， LUKAN J， et al. A deep learning model for burn depth classification using ultrasound imaging［J］. J Mech Behav Biomed Mater， 2022， 125： 104930.
[16]	DOBRE E G， SURCEL M， CONSTANTIN C， et al. Skin cancer pathobiology at a glance： a focus on imaging techniques and their potential for improved diagnosis and surveillance in clinical cohorts［J］. Int J Mol Sci， 2023， 24（2）： 1079.
[17]	ALHEEJAWI S， BERENDT R， JHA N， et al. Detection of malignant melanoma in H&E-stained images using deep learning techniques［J］. Tissue Cell， 2021， 73： 101659.
[18]	ZHAO B， CHEN X， LI Z， et al. Triple U-net： hematoxylin-aware nuclei segmentation with progressive dense feature aggregation［J］. Med Image Anal， 2020， 65： 101786.
[19]	WANG C， MA Q， WEI Y， et al. Deep learning automatically assesses 2-µm laser-induced skin damage OCT images［J］. Lasers Med Sci， 2024， 39（1）： 106.
[20]	GROH M， BADRI O， DANESHJOU R， et al. Deep learning-aided decision support for diagnosis of skin disease across skin tones［J］. Nat Med， 2024， 30（2）： 573-583.
[21]	MAHBOD A， SCHAEFER G， WANG C， et al. Transfer learning using a multi-scale and multi-network ensemble for skin lesion classification［J］. Comput Methods Programs Biomed， 2020， 193： 105475.
[22]	TANG P， YAN X， NAN Y， et al. FusionM4Net： A multi-stage multi-modal learning algorithm for multi-label skin lesion classification［J］. Med Image Anal， 2022， 76： 102307.
[23]	WANG Y， FENG Y， ZHANG L， et al. Adversarial multimodal fusion with attention mechanism for skin lesion classification using clinical and dermoscopic images［J］. Med Image Anal， 2022， 81： 102535.
[24]	LEE K， CAVALCANTI T C， KIM S， et al. Multi-task and few-shot learning-based fully automatic deep learning platform for mobile diagnosis of skin diseases［J］. IEEE J Biomed Health Inform， 2023， 27（1）： 176-187.
[25]	ALDHYANI T H H， VERMA A， AL-ADHAILEH M H， et al. Multi-class skin lesion classification using a lightweight dynamic kernel deep-learning-based convolutional neural network［J］. Diagnostics （Basel）， 2022， 12（9）： 2048.
[26]	RAGHAVENDRA P V S P， CHARITHA C， BEGUM K G， et al. Deep learning-based skin lesion multi-class classification with global average pooling improvement［J］. J Digit Imaging， 2023， 36（5）： 2227-2248.
[27]	PACHECO A G C， KROHLING R A. An attention-based mechanism to combine images and metadata in deep learning models applied to skin cancer classification ［J］. IEEE J Biomed Health Inform， 2021， 25（9）： 3554-3563.
[28]	THIEME A H， ZHENG Y， MACHIRAJU G， et al. A deep-learning algorithm to classify skin lesions from mpox virus infection［J］. Nat Med， 2023， 29（3）： 738-747.
[29]	NINGRUM D N A， YUAN S P， KUNG W M， et al. Deep learning classifier with patient’s metadata of dermoscopic images in malignant melanoma detection［J］. J Multidiscip Healthc， 2021， 14： 877-885.
[30]	PACHECO A G C， KROHLING R A. The impact of patient clinical information on automated skin cancer detection［J］. Comput Biol Med， 2020， 116： 103545.
[31]	CAI G， ZHU Y， WU Y， et al. A multimodal transformer to fuse images and metadata for skin disease classification［J］. Vis Comput， 2023， 39（7）：2781-93.
[32]	COMBALIA M， CODELLA N， ROTEMBERG V， et al. Validation of artificial intelligence prediction models for skin cancer diagnosis using dermoscopy images： the 2019 International Skin Imaging Collaboration Grand Challenge［J］. Lancet Digit Health， 2022， 4（5）： e330-e339.
[33]	AGHAMOHAMMADESMAEILKETABFOROOSH K， PARFITT J， NIKAN S， et al. From blender to farm： transforming controlled environment agriculture with synthetic data and SwinUNet for precision crop monitoring ［J］. PLoS One， 2025， 20（4）： e0322189.
[34]	FUJISAWA Y， OTOMO Y， OGATA Y， et al. Deep-learning-based， computer-aided classifier developed with a small dataset of clinical images surpasses board-certified dermatologists in skin tumour diagnosis ［J］. Br J Dermatol， 2019， 180（2）： 373-381.
[35]	ALAM T M， SHAUKAT K， KHAN W A， et al. An efficient deep learning-based skin cancer classifier for an imbalanced dataset［J］. Diagnostics （Basel）， 2022， 12（9）： 2115.
[36]	ZHU A Q， WANG Q， SHI Y L， et al.； China Alliance of Multi-Center Clinical Study for Ultrasound （Ultra-Chance）. A deep learning fusion network trained with clinical and high-frequency ultrasound images in the multi-classification of skin diseases in comparison with dermatologists： a prospective and multicenter study ［J］. EClinicalMedicine， 2024， 67： 102391.
[37]	XIN C， LIU Z， ZHAO K， et al. An improved transformer network for skin cancer classification［J］. Comput Biol Med， 2022， 149： 105939.
[38]	THOMAS S M， LEFEVRE J G， BAXTER G， et al. Interpretable deep learning systems for multi-class segmentation and classification of non-melanoma skin cancer［J］. Med Image Anal， 2021， 68： 101915.
[39]	LUCIERI A， BAJWA M N， BRAUN S A， et al. ExAID： A multimodal explanation framework for computer-aided diagnosis of skin lesions［J］. Comput Methods Programs Biomed， 2022， 215： 106620.
[40]	RASHEED A， UMAR A I， SHIRAZI S H， et al. Automatic eczema classification in clinical images based on hybrid deep neural network［J］. Comput Biol Med， 2022， 147： 105807.
[41]	HAMMAD M， PŁAWIAK P， ELAFFENDI M， et al. Enhanced deep learning approach for accurate eczema and psoriasis skin detection ［J］. Sensors （Basel）， 2023， 23（16）： 7295.
[42]	FOLLE L， FENZL P， FAGNI F， et al. DeepNAPSI multi-reader nail psoriasis prediction using deep learning［J］. Sci Rep， 2023， 13（1）： 5329.
[43]	YANG Y， GUO L， WU Q， et al. Construction and evaluation of a deep learning model for assessing acne vulgaris using clinical images［J］. Dermatol Ther （Heidelb）， 2021， 11（4）： 1239-1248.
[44]	GUO L， YANG Y， DING H， et al. A deep learning-based hybrid artificial intelligence model for the detection and severity assessment of vitiligo lesions［J］. Ann Transl Med， 2022， 10（10）： 590.
[45]	ZHENG X， HUANG Y， LIU W， et al. LW-XNet for segmentation and classification of skin lesions from dermoscopy images［J］. Expert Syst Appl， 2024， 255： 124826.
[46]	WANG L， SHAO A， HUANG F， et al. Deep learning-based semantic segmentation of non-melanocytic skin tumors in whole-slide histopathological images［J］. Exp Dermatol， 2023， 32（6）： 831-839.
[47]	DE LOGU F， UGOLINI F， MAIO V， et al. Recognition of cutaneous melanoma on digitized histopathological slides via artificial intelligence algorithm［J］. Front Oncol， 2020， 10： 1559.
[48]	AGGARWAL P， PAPAY F A. Artificial intelligence image recognition of melanoma and basal cell carcinoma in racially diverse populations［J］. J Dermatolog Treat， 2022， 33（4）： 2257-2262.
[49]	PHAM T C， LUONG C M， HOANG V D， et al. AI outperformed every dermatologist in dermoscopic melanoma diagnosis， using an optimized deep-CNN architecture with custom mini-batch logic and loss function［J］. Sci Rep， 2021， 11（1）： 17485.
[50]	KIM R H， NOMIKOU S， COUDRAY N， et al. Deep learning and pathomics analyses reveal cell nuclei as important features for mutation prediction of BRAF-mutated melanomas［J］. J Invest Dermatol， 2022， 142（6）： 1650-1658.
[51]	HEKLER A， UTIKAL J S， ENK A H， et al. Pathologist-level classification of histopathological melanoma images with deep neural networks［J］. Eur J Cancer， 2019， 115： 79-83.
[52]	THOMSEN K， CHRISTENSEN A L， IVERSEN L， et al. Deep learning for diagnostic binary classification of multiple-lesion skin diseases ［J］. Front Med （Lausanne）， 2020， 7： 574329
[53]	LIU Y， JAIN A， ENG C， et al. A deep learning system for differential diagnosis of skin diseases ［J］. Nat Med， 2020， 26（6）： 900-908.

皮肤病深度学习诊断模型的研究进展

Research progress in deep learning-based diagnostic models for dermatological diseases

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

参考文献 53

相关文章 2

Metrics

本文评价

推荐阅读 0

[1]	白雪,王晨晨,石张镇,毕林涛. 多发性骨髓瘤患者胸部CT第4胸椎平面人体成分与预后的关联性分析[J]. 吉林大学学报(医学版), 2024, 50(4): 1098-1108.
[2]	钟淑霞, 李珊山, 姚蕾, 刘鹤松, 田亚萍, 温泉. 长脉冲可调脉宽1 064 nm Nd∶YAG激光在血管性皮肤病治疗中的临床应用[J]. J4, 2011, 37(1): 147-149.