Investigating the correlation between acrylamide content and browning ratio of model cookies

A computer vision based image analysis algorithm was developed for the prediction of acrylamide level in cookies. Digital images of cookies were segmented based on the predefined color references representing the mean red, green and blue values of brownish and cream-colored regions. The browning ratio was defined as the ratio of brown pixels to total pixels and calculated using the segmented images of cookies. Cookies were prepared from four recipes by baking at 200 and 220 °C for different times up to 25 min. The rates of browning and acrylamide formation followed almost the same kinetic pattern during baking. Recipe formulation and baking conditions affected these rates in a similar way. A significant correlation (r = 0.946, p < 0.01) was observed between the browning ratio and acrylamide concentration in cookies. A browning ratio of less than 8% indicated that acrylamide concentration is below its detection limits determined by liquid chromatography–mass spectrometry. The success of the algorithm was 100% on sorting cookies based on a threshold level of 150 ng/g acrylamide.