Systematic Review of Deep Learning and Machine Learning Models for Eczema (Atopic Dermatitis) Image Detection
Bakhtar Journal of Engineering and Technology · 30.12.2025 · 1 (2) · 35
DOI: 10.65460/bjet/vol1_iss2_a4 ↗Kurz gefasst
Systematische Übersicht in der Zeitschrift Bakhtar Journal of Engineering and Technology, online erschienen am 30.12.2025. Verfasst von Ishan Safi und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Atopic dermatitis (eczema) is a chronic inflammatory skin disease with highly variable visual manifestations, posing significant challenges for accurate diagnosis and severity assessment in clinical practice. In recent years, artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL) techniques, has demonstrated strong potential for automated dermatological image analysis. However, existing studies on AIbased eczema detection exhibit substantial variability in model architectures, datasets, evaluation protocols, and clinical objectives, limiting reproducibility and real-world deployment. This paper presents a systematic review of machine learning and deep learning models applied to eczema image analysis, covering studies published between 2020 and 2025. Following the PRISMA framework, selected studies are analyzed with respect to model categories, including classical ML methods, convolutional neural networks (CNNs), Vision Transformers (ViTs), segmentation-based architectures, and hybrid ML–DL approaches. Performance trends, dataset characteristics, preprocessing strategies, and evaluation metrics are critically examined. The findings indicate that deep learning models particularly CNN-based, attention-enhanced, and hybrid architecture consistently outperform traditional ML methods in classification and severity-related tasks. Segmentationbased approaches further enable objective lesion localization and severity estimation, while hybrid models offer improved stability in data-limited settings. Nevertheless, challenges related to dataset scarcity, generalization across skin tones, lack of standardized benchmarks, and limited explainability remain significant barriers to clinical adoption. This review provides a structured taxonomy of existing approaches, highlights key research gaps, and offers practical recommendations to guide future research toward reliable, interpretable, and clinically applicable AI-driven eczema detection systems.
Zitieren
APA
Safi, I., Safi, F. M., & Pashay, R. (2025). Systematic Review of Deep Learning and Machine Learning Models for Eczema (Atopic Dermatitis) Image Detection. Bakhtar Journal of Engineering and Technology, 1(2), 35. https://doi.org/10.65460/bjet/vol1_iss2_a4
Vancouver
Safi I, Safi FM, Pashay R. Systematic Review of Deep Learning and Machine Learning Models for Eczema (Atopic Dermatitis) Image Detection. Bakhtar Journal of Engineering and Technology. 2025;1(2):35. doi:10.65460/bjet/vol1_iss2_a4
BibTeX
@article{Safi2025,
author = {Safi, Ishan and Safi, Fida Mohammad and Pashay, Rahimullah},
title = {Systematic Review of Deep Learning and Machine Learning Models for Eczema (Atopic Dermatitis) Image Detection},
journal = {Bakhtar Journal of Engineering and Technology},
year = {2025},
volume = {1},
number = {2},
pages = {35},
doi = {10.65460/bjet/vol1_iss2_a4}
}
EUVIMED