Diabetes Foot Ulcer Diagnosis using Fast Convolution Neural Network

Diabetic foot ulcer (DFU) is a major health concern for people with diabetes. The physician’s quality of life suffers, the health care system incurs costs, and amputation may be necessary if these wounds are not treated promptly. Automated detection methods might aid medical professionals in the fight against and treatment of disease. Several machine learning-based approaches to fixing this problem have been proposed recently. Specifically, the study recommends utilizing deep learning techniques to spot ulcers in patient images of their feet; this might aid in the diagnosis and management of diabetic foot ulcers. To make the current Fast Convolution Neural Network (FCNN) better, we propose enhancing it using data augmentation techniques and new parameter settings. They used 2000 DFU images from a properly labeled training dataset. The training was verified using Monte Carlo cross-validation. Standard detector implementations were outperformed by our method’s average mean accuracy of 92.9%, Fl-score of 93.2%, and average recognition speed of l00.8ms.