<p>Automatic detection and grading of diabetic foot ulcers is necessary for effective intervention and therapy. Categorizing DFUs requires careful experimentation and hyperparameter tuning. The model went through rigorous training and testing phases using a dataset that included images from various hospitals. We perform key procedures such as data preparation, annotation, and augmentation to address the imbalance problem. Model performance was improved by systematically adjusting parameters such as epochs, learning rates, batch sizes, and network architectures. The presented models performed well in classifying DFUs. This shows the efficacy of the proposed approaches and architectures. This study emphasizes the potential for practical application in medical image analysis, as well as the importance of hyperparameter tuning in deep learning model development. The proposed model performs excellent detection and grading with 100% testing accuracy on the dataset. The findings of this study could improve diabetic foot ulcer management, allowing for more effective therapies. Future research will look into merging the model with telemedicine for remote diabetic foot ulcer care.</p>