Efficient Net-Based Deep Learning Model for Real-Time Diabetic Foot Ulcer Detection

Abstract

Diabetic foot ulcers (DFUs) are among the most serious complications of diabetes and can lead to severe infections, tissue damage, and amputations if not detected early. Early and accurate diagnosis is essential for improving patient outcomes and reducing healthcare risks. Traditional methods of DFU diagnosis mainly depend on manual examination by healthcare professionals, which can be time-consuming, subjective, and inconsistent. To address these challenges, this research proposes an automated diabetic foot ulcer detection system using the EfficientNet deep learning model. The proposed model utilizes transfer learning techniques to extract important and discriminative features from medical foot images for accurate classification of ulcer and non-ulcer cases. Several preprocessing techniques, including denoising, normalization, and data augmentation methods such as rotation, flipping, and scaling, are applied to improve image quality and reduce overfitting. Experimental results show that the EfficientNet-based model achieves a high detection accuracy of 97.5%, demonstrating its effectiveness and reliability in medical image classification. proposed approach can serve as a valuable decision-support tool for healthcare professionals by enabling early diagnosis and better management of diabetic foot ulcers. This system has the potential to improve clinical efficiency and reduce complications associated with delayed detection of DFUs.