Narrow Results

Pathological analysis as well as biomechanical methods are powerful approaches for collagen assessment, which plays an important role in understanding the wound healing process and choosing a treatment method in clinical situations. Due to the limitations of preparing and evaluating pathological images, this study was designed to establish a machine learning technique to predict the wound collagen content through its biomechanical parameters. For this purpose, the artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) were compared. The wound was created with an incision on the back of 30 male BALB/c mice. On the 7th and 14th days, animals were sacrificed and 60 wound tissue samples were evaluated using histopathological and biomechanical methods to quantify the amount of collagen and wound tensile strength to feed the ANN and ANFIS developed models. Based on the results, both models have appropriate performance to predict the wound collagen content. However, the comparison of coefficient of determination ($R^2$) and root mean square error (RMSE) for testing dataset revealed that ANN ($R^2=0.95$, $\mathrm{\text{RMSE}}=0.29$) had more prediction capability than ANFIS ($R^2=0.84$, $\mathrm{\text{RMSE}}=0.87$). As a decision support system, ANN model could assist in the evaluation of wound healing process with collagen values prediction.