A Review on Deep Learning Techniques for Diabetic Foot Ulcer Classification

Diabetes is common nowadays and hyperglycaemia index is high then it is the case of diabetes. If there is a wound on a leg and it is for many years not cured then it may be dangerous. In this case there are high risk of diabetic foot ulcer. If this is not treated at right time then the patient may suffer from severe infection or amputation. To prevent from foot diabetic many developments has been made for early prediction using artificial intelligence. Deep learning models have widely used in processing the medical image dataset that have achieved higher performance. Deep learning made easy to classify severity of foot ulcer within less time and helps patient to know about the diseases at early stage and because of which patient can get right treatment. Automated analysis of foot images, including photos and thermal images are used for processing. DL models are capable of precisely identifying areas of healthy skin from those with ulcers, determining the extent and size of wounds, and tracking their healing over time. With live dataset we can achieve better results for ulcer presentation. DFU images are used as a dataset for predicting the system but if clinical reports are also used then it may help in giving higher performance in terms of accuracy and efficiency. Patients who are more likely to experience DFUs can be identified by predictive models, enabling proactive treatment and focused interventions. This paper consist of review of different techniques used till now for classification of ulcer and a proposed framework to identify ulcer severity at early stage.