Diabetic Foot Ulcer using Machine Learning

Diabetic foot ulcers (DFUs) are a significant and life-threatening complication of diabetes, affecting 15% to 25% of diabetic individuals worldwide. If not diagnosed and managed early, DFUs can lead to severe infections, prolonged hospitalization, and lower limb amputations. Traditional diagnostic methods on clinical examination and subjective assessment, may result in delayed intervention. This study presents a machine learning-based framework for the automated detection and classification of DFUs to assist in early diagnosis and timely treatment. A dataset of 1,055 images was used, comprising 512 normal (healthy) foot images and 543 abnormal (DFU-affected) images, sourced from Kaggle. The study evaluates four machine learning models Support Vector Machines (SVM), Random Forest (RF), Logistic Regression (LR), and AdaBoost to classify images into normal and DFU categories. Performance evaluation was conducted using key metrics such as accuracy, precision, recall, F1-score, and specificity. The results demonstrate that SVM and LR achieved the highest accuracy of 98%, outperforming RF and AdaBoost. These findings suggest that machine learning can serve as an efficient and reliable diagnostic tool, reducing the dependency on manual assessment and enabling faster, more objective, and cost-effective DFU detection. Future work will focus on improving model generalization using larger datasets and advanced deep-learning techniques.