What question did this study set out to answer?

To enhance early detection of diabetic neuropathy using an explainable deep learning framework.

April 5, 2026Scientific ReportsOpen Access

SoleFusion-Net: an explainable multimodal deep learning framework for diabetic foot syndrome classification in type II diabetes mellitus

View Full Paper

Ask AI

Bookmark

Authors

This framework demonstrates classification of diabetic foot syndrome using multimodal data, suggesting improved detection methods.

To enhance early detection of diabetic neuropathy using an explainable deep learning framework.
Developed SoleFusion-Net integrating plantar pressure images and clinical data
Employed a late-fusion dual-branch architecture
Utilized convolutional layers for image analysis and tabular processing for clinical variables
Implemented explainability techniques like Grad-CAM and SHAP
Evaluated framework with 504 patients categorized into neuropathy severity levels.
Achieved 83% validation accuracy
AUCs of 0.962, 0.892, and 0.933 for mild, moderate, and severe neuropathy respectively
Identified vibration perception threshold and monofilament scores as key predictors
Demonstrated explainability through spatial analysis using Grad-CAM.

Sheikh et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark