Optimized Fuzzy Logic and Adaptive Neuro-Fuzzy Inference Systems for Wound Healing Time Prediction Among the Diabetic Patients

Compared with normal wounds, chronic diabetes wounds are considered more severe because healing takes a significant amount of time. Among the factors that affect the healing of chronic wounds, infection is a major factor that slows down wound healing. Until now, medical practitioners assess wound infection via visual observation, followed by a culture of wound tissue. However, this traditional type of wound assessment is complex and not readily accessible, so the utilization of fuzzy logic to solve the problem mentioned above is proposed. Systematic literature review and medical expert consultation were utilized in determining the various biomarkers that affect the healing period of a wound, constructing the ranges for membership functions and working rules of each of these biomarkers, and developing prediction healing time. Based on pH, oxygen level, c-reactive protein (CRP), and vascular endothelial growth factor (VEGF), the fuzzy logic (FL) system and the adaptive neuro-fuzzy inference system (ANFIS) models were developed through MATLAB software. A system consisting of both models was constructed to verify of each model. Comparing the results, the model based on fuzzy logic showed a higher accuracy of 95%. This developed model can be applied in developing the smart wound monitoring system in the future.