Fractal colour: a new approach for evaluation of acrylamide contents in biscuits

A new method to evaluate the content of acrylamide (AA) in biscuits was proposed. Principal component analysis (PCA) and least-squares support vector machine (LS-SVM) combined with fractal colour were developed and compared with the model based on other colour spaces, such as RGB, CMYK and L(*)a(*)b(*). The results of this study show fractal colour has the best performance for the classification of the biscuits contained different levels of AA by PCA. In addition, the prediction capacity of LS-SVM based on fractal colour [root mean square error (RMSE)=15.70 ng/g, coefficient of determination (R(2))=0.99] is much better than that based on RGB (RMSE=26.08 ng/g, R(2)=0.98) and CMYK (RMSE=17.99 ng/g, R(2)=0.99), though worse than L(*)a(*)b(*) (RMSE=7.56 ng/g, R(2)=1.00). Therefore, fractal colour could be a new colour feature to characterise the colour information for the evaluation of AA content in biscuits.