Related Experiment Video
Updated: May 22, 2025
13:19
Deep Neural Networks for Image-Based Dietary Assessment
Published on: March 13, 2021
8.9K
An Integrated Deep Learning Model with EfficientNet and ResNet for Accurate Multi-Class Skin Disease Classification
Madallah Alruwaili1, Mahmood Mohamed2
1Department of Computer Engineering and Networks, College of Computer and Information Sciences, Jouf University, Sakaka 72388, Aljouf, Saudi Arabia.
Diagnostics (Basel, Switzerland)
|March 13, 2025
Summary
A novel deep learning model combining three CNNs achieves 99.14% accuracy for skin disease classification. This fusion-level approach enhances diagnostic stability and precision for conditions like skin cancer.
Area of Science:
- Dermatology
- Artificial Intelligence
- Medical Imaging
Background:
- Medical diagnosis of skin diseases is challenging due to patient variations.
- Accurate classification of conditions like skin cancer and neoplasms is critical.
Purpose of the Study:
- To develop a stable, high-performance deep learning model for skin disease classification.
- To improve diagnostic accuracy through a fusion-level approach.
Main Methods:
- A fusion-level deep learning model merging EfficientNet-B0, EfficientNet-B2, and ResNet50 was designed.
- The model extracts features using distinct CNN branches and a fusion mechanism.
- The Kaggle Skin Diseases Image Dataset (27,153 images) was used for training, validation, and testing.
Main Results:
- The proposed model achieved 99.14% accuracy.
- Excellent precision, recall, and F1-score metrics were obtained.
- The model demonstrated high classification precision.
Conclusions:
- The deep learning model shows significant potential for automated dermatological diagnosis.
- The approach is promising for clinical applications in skin disease classification.
