Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Abstract. Cooking loss (CL), which is one of the many diverse methods for water-holding capacity (WHC) determination, is an important quality trait directly relating to the eating quality of broiler breast fillets. In this study, hyperspectral imaging (HSI) in short-wave infrared (SWIR, 1000-2500 nm) range was investigated for non-invasive assessment of CL of fresh broiler breast fillets. A total of 74 SWIR hyperspectral images for fillets were acquired, and after images calibration, principal component analysis (PCA) followed by inverse PCA was conducted to eliminate bad lines and strip noise. Mean spectra were extracted from regions of interest (ROIs) identified by a mask. Fillets CL were determined by measuring the weight loss in cooking and CL categories were defined as high-CL when CL ≥ 20%, and low-CL when CL ˂ 20%. Two classification methods of partial least square discriminant analysis (PLS-DA) and radial basis function-support vector machine (RBF-SVM) were applied, respectively, with various preprocessing methods including first or second derivative (Der1 and Der2), standard normal variate (SNV) and multiplicative scatter correction (MSC). Overall, RBF-SVM model with the optimal preprocessing of MSC was considered as the best model by showing correct classification rates (CCRs) of 0.78 and 0.71 for calibration and prediction sets, respectively. Conclusively, results suggest that SWIR HSI has limited benefits for assessing CL of fresh broiler breast fillets; however, the classification result should be enhanced in further study.