RGB pattern of images allows rapid and efficient prediction of antioxidant potential in Calycophyllum spruceanum barks

Abstract

The use of fast and low-cost methods to optimize the total phenolic compounds (TPC) extraction has been gaining attention in ethnopharmacological research. Extraction conditions of the bioactive compounds from Calycophyllum spruceanum barks were established through multivariate regression models. In this sense, fractional factorial design (FFD) and central rotational composite design (CCRD) were developed using partial least squares regression (PLSR) combined with the information from the color images and spectrophotometry tools to evaluate the antioxidant activity from C. spruceanum barks. In fact, was possible to optimize the extraction of TPC with AA (ethanol 10% v/v, 1 h extraction time at 75 °C temperature). Besides, the precision and performance of generated models were established for the three response variables (TPC, AA by ABTS and FRAP methods) with R2 above 0.98 in the PLSR and residual predictive value (RPD) above 3. Thus, the approaches suggested in this study, with emphasis on the use of image analysis, proved to be potential and promising as simple, fast, non-destructive methods for quantifying TPC and antioxidant activity in C. spruceanum barks.

Keywords

• Calycophyllum spruceanum
• Antioxidant potential
• Optimization
• Chemometric
• Computer vision
• Image analysis