Predicting mechanical properties of fried chicken nuggets using image processing and neural network techniques

Typical approaches for measuring mechanical properties of fried food products are mostly destructive techniques. In this study, a non-destructive, image-based method was evaluated for predicting mechanical properties of fried, breaded chicken nuggets. The textural parameters of interest, namely maximum load, energy to break point, and toughness of fried chicken nuggets were measured. Values of the parameters changed over frying time. Images of the chicken nuggets were collected at different frying stages and five image texture indices were extracted using co-occurrence matrix. A multiple-layer feed-forward neural network was established to predict the three mechanical parameters. The correlation coefficients of the predicted results with those from mechanical tests were above 0.84.