EfficientNet B3-Based Deep Learning-Based Diabetic Foot Ulcer Classification: A Study Using Thermal Images

Diabetic foot ulcers diseases pose a significant challenge to diabetes care hence requires rapid accurate diagnosis to enhance treatment. This research explores the application of efficient nets B3 deep learning model in the categorization of thermal images for classification of diabetic foot ulcers. The research encompasses a comprehensive dataset that has been divided into two distinct sets: They divided the set into a training set, being 80% of the set, and test set being 20% of the set. In total, training of the model has occurred in 15 epochs. Our findings confirm a good degree of success in the process of classification as applied in the subject of the research. In particular, it has been observed from the proposed EfficientNet B3 model that it possesses the outstanding accuracy of nearly 99.66% when tested using the testing dataset. having shown in the paper that it is possible to correctly distinguish and classify diabetic foot ulcers into various stages, the model could be seen as a feasible tool in clinical settings. It makes to raise the quality of lifes.