DFU-Helper: An Innovative Framework for Longitudinal Diabetic Foot Ulcer Diseases Evaluation Using Deep Learning

Diabetes affects roughly 537 million people, and is predicted to reach 783 by 2045. Foot Ulcer (DFU) a major complication associated with diabetes can lead lower limb amputation. The rapid evolution of diabetic foot ulcers (DFUs) necessitates immediate intervention prevent the severe consequences amputation related complications. Continuous meticulous patient monitoring for individuals crucial currently carried out medical practitioners on daily basis. This research article introduces DFU-Helper, novel framework that employs Siamese Neural Network (SNN) accurate objective assessment progression over time. DFU-Helper provides healthcare professionals comprehensive visual numerical representation in terms similarity distance disease, considering five distinct disease conditions: none, infection, ischemia, both (presence ischemia infection), healthy. SNN achieves best Macro F1-score 0.6455 test dataset when applying pseudo-labeling pseudo-threshold set 0.9. used process creating anchors each class using feature vectors. When initially consults professional, an image transmitted model, which computes distances from anchor point. It generates table corresponding figures visually intuitive radar chart. In subsequent visits, another captured fed into model alongside initial image. then plots images presents points. Our proposed system represents significant advancement application deep learning longitudinal DFU. To our knowledge, no existing tool harnesses DFU follow-up comparable manner.