XMP-Net: An XAI-Based Modified Xception Model for Recognizing Monkeypox and Other Skin Diseases

Area of Science:

• Dermatology
• Artificial Intelligence
• Medical Imaging

Background:

• Accurate and timely diagnosis of skin conditions, including monkeypox, is crucial for public health.
• Traditional diagnostic methods can be time-consuming and may require specialized expertise.
• The integration of artificial intelligence (AI) offers potential for improved diagnostic accuracy and efficiency.

Purpose of the Study:

• To develop and evaluate a deep learning model (XMP-Net) for classifying skin conditions: normal, chickenpox, measles, and monkeypox.
• To enhance the interpretability of the AI model's predictions using Grad-CAM and LIME.
• To assess the model's performance in terms of accuracy, precision, recall, and F1-score, particularly for monkeypox detection.

Main Methods:

• A modified Xception-based deep learning architecture, termed XMP-Net, was designed for skin image classification.
• Grad-CAM (gradient-weighted class activation mapping) and LIME (local interpretable model-agnostic explanations) were utilized for model interpretability.
• The model was trained and validated on a dataset comprising images of normal skin, chickenpox, measles, and monkeypox.

Main Results:

• XMP-Net achieved high classification accuracy: 98.33% for normal skin, 98.25% for monkeypox, 84.21% for measles, and 77.27% for chickenpox.
• For monkeypox, the model demonstrated a precision of 91.80%, recall of 98.25%, and an F1-score of 94.92%.
• Visual explanations from Grad-CAM and LIME identified key image regions influencing predictions, offering clinical insights.

Conclusions:

• XMP-Net shows significant potential for accurate and interpretable skin condition diagnosis, especially for monkeypox.
• Explainable AI (XAI) techniques can increase confidence in AI-driven diagnostic tools.
• The developed model serves as a foundation for accessible, reliable diagnostic tools, particularly in resource-limited settings.