A Secure Framework toward IoMT-Assisted Data Collection, Modeling, and Classification for Intelligent Dermatology...

Area of Science:

• Dermatology
• Medical Imaging
• Artificial Intelligence

Background:

• Skin cancer, including basal cell carcinoma, squamous cell carcinoma, and melanoma, is a significant dermatological concern.
• Early diagnosis of skin lesions is crucial for effective treatment and patient outcomes.
• Remote screening methods are needed to improve accessibility and timely detection.

Purpose of the Study:

• To develop and evaluate a secure Internet of Medical Things (IoMT)-assisted framework for remote initial screening of skin lesions.
• To enhance the accuracy and efficiency of skin lesion classification using an ensemble deep learning approach.

Main Methods:

• An IoMT-based device was used for patients to capture and securely transmit encrypted skin lesion images to cloud storage.
• An ensemble classifier combining VGG-16, DenseNet-201, Inception-V3, and Efficient-B7 Convolutional Neural Network (CNN) models was employed for image classification.
• The framework was validated using the "HAM10000" dataset, comprising 7 different types of skin lesions.

Main Results:

• The ensemble model achieved the highest accuracy at 87.22%, outperforming individual models like DenseNet-201.
• The ensemble model demonstrated a lower test loss/error of 0.4131.
• A high Area Under the Curve (AUC) score of 0.9745 indicated strong classification ability.

Conclusions:

• The proposed IoMT-assisted framework provides a secure and effective solution for remote skin lesion screening.
• Ensemble deep learning models significantly improve the accuracy of skin cancer diagnosis compared to single models.
• The framework supports timely patient assessment and recommendations by a medical team based on Computer-Aided Diagnosis (CAD) results.