Detection of diabetic foot ulcer (DFU) with AlexNet and ResNet-101

Diabetic foot ulcer (DFU) is a major complication of diabetes mellitus which can result into lower limb elimination if not managed appropriately. The clinical detection of DFU is challenging due to the duration and high cost. Thus, computerized methods are needed to overcome these challenges. We proposed the use of AlexNet and ResNet-101 models for the classification of foot images to detect DFU. A dataset of foot images containing DFU and healthy skin was obtained from the Kaggle database. Different dataset split ratio and epoch were employed to evaluate the optimum performance of these models in this study. The best results were attained at 80:20 split and 40 epoch. AlexNet attained 97.1% accuracy, 100% sensitivity and 91.7% specificity, while ResNet attained 97.1% accuracy, 95.7% sensitivity and 100% specificity. These models showed good results and will be helpful in DFU diagnosis.