Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Implementation of a Skin Lesion Detection System for Managing Children with Atopic Dermatitis Based on Ensemble Learning

Soobin Jeon, Sujong Kim, Dongmahn Seo

TL;DR

This work introduces ENSEL, an ensemble learning–based skin lesion detection system designed to improve objective diagnosis of pediatric atopic dermatitis within real-world clinical data. By integrating Mask R-CNN for localization, YOLOv8 for fast detection, and EfficientNet for robust classification, and fusing predictions with soft voting, ENSEL achieves high recall and competitive speed on real-world, noisy images. The authors validate ENSEL against single-model baselines on a Korean real-world dataset, reporting recall around 0.98 and a majority of results under 1 second, with Grad-CAM visualizations enhancing interpretability. The study demonstrates the practicality of cloud-based, data-driven skin-disease management tools and outlines future commercialization and optimization directions.

Abstract

The amendments made to the Data 3 Act and impact of COVID-19 have fostered the growth of digital healthcare market and promoted the use of medical data in artificial intelligence in South Korea. Atopic dermatitis, a chronic inflammatory skin disease, is diagnosed via subjective evaluations without using objective diagnostic methods, thereby increasing the risk of misdiagnosis. It is also similar to psoriasis in appearance, further complicating its accurate diagnosis. Existing studies on skin diseases have used high-quality dermoscopic image datasets, but such high-quality images cannot be obtained in actual clinical settings. Moreover, existing systems must ensure accuracy and fast response times. To this end, an ensemble learning-based skin lesion detection system (ENSEL) was proposed herein. ENSEL enhanced diagnostic accuracy by integrating various deep learning models via an ensemble approach. Its performance was verified by conducting skin lesion detection experiments using images of skin lesions taken by actual users. Its accuracy and response time were measured using randomly sampled skin disease images. Results revealed that ENSEL achieved high recall in most images and less than 1s s processing speed. This study contributes to the objective diagnosis of skin lesions and promotes the advancement of digital healthcare.

Implementation of a Skin Lesion Detection System for Managing Children with Atopic Dermatitis Based on Ensemble Learning

TL;DR

This work introduces ENSEL, an ensemble learning–based skin lesion detection system designed to improve objective diagnosis of pediatric atopic dermatitis within real-world clinical data. By integrating Mask R-CNN for localization, YOLOv8 for fast detection, and EfficientNet for robust classification, and fusing predictions with soft voting, ENSEL achieves high recall and competitive speed on real-world, noisy images. The authors validate ENSEL against single-model baselines on a Korean real-world dataset, reporting recall around 0.98 and a majority of results under 1 second, with Grad-CAM visualizations enhancing interpretability. The study demonstrates the practicality of cloud-based, data-driven skin-disease management tools and outlines future commercialization and optimization directions.

Abstract

The amendments made to the Data 3 Act and impact of COVID-19 have fostered the growth of digital healthcare market and promoted the use of medical data in artificial intelligence in South Korea. Atopic dermatitis, a chronic inflammatory skin disease, is diagnosed via subjective evaluations without using objective diagnostic methods, thereby increasing the risk of misdiagnosis. It is also similar to psoriasis in appearance, further complicating its accurate diagnosis. Existing studies on skin diseases have used high-quality dermoscopic image datasets, but such high-quality images cannot be obtained in actual clinical settings. Moreover, existing systems must ensure accuracy and fast response times. To this end, an ensemble learning-based skin lesion detection system (ENSEL) was proposed herein. ENSEL enhanced diagnostic accuracy by integrating various deep learning models via an ensemble approach. Its performance was verified by conducting skin lesion detection experiments using images of skin lesions taken by actual users. Its accuracy and response time were measured using randomly sampled skin disease images. Results revealed that ENSEL achieved high recall in most images and less than 1s s processing speed. This study contributes to the objective diagnosis of skin lesions and promotes the advancement of digital healthcare.

Paper Structure

This paper contains 20 sections, 14 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Commercial Software for Skin Disease Detection
  4. Conventional Methodologies for Skin Disease Detection
  5. Real-World Dataset for Skin Disease Detection
  6. Proposed System Architecture
  7. Data Exchange Module
  8. Diagnostic Result Analysis Module
  9. Skin Lesion Inference Module
  10. Deep Learning Model Training
  11. Evaluation of Deep Learning Model Performance
  12. Application of Ensemble Methods
  13. System Implementation and Experimental Results
  14. System Environment
  15. System Implementation
  16. ...and 5 more sections

Figures (14)

  • Figure 1: Architecture of the ensemble learning–based skin lesion detection system
  • Figure 2: Diagnostic flowchart of ENSEL for detecting skin diseases
  • Figure 3: Example of Grad-CAM visualization applied to skin disease images
  • Figure 4: Diagnostic flowchart of the skin lesion inference module
  • Figure 5: Mask R-CNN box regression loss graph: (a) train box regression loss and (b) validation regression loss.
  • ...and 9 more figures