Predicting Color Traits of Intact Broiler Breast Fillets Using Visible and Near-Infrared Spectroscopy

This study was conducted to assess the potential for visible and near-infrared (Vis/NIR) spectroscopy to predict an important quality attribute, color traits, of intact broiler breast fillets (pectoralis major). A total of 214 boneless and skinless chicken fillets were procured from a commercial processing plant. The quantitative calibration models were established between five color trait parameters (CIELAB L*, a*, b*, chroma, and hue angle) and spectra from four different spectral ranges (400-2500, 400-1100, 1100-2500, and 900-1700 nm) collected from skin (ventral) and bone (dorsal) sides of fillets individually by partial least squares regressions (PLSR). Predictive ability was assessed by coefficient of determination of prediction (R (p) (2)), ratio performance deviation (RPD), and range error ratio (RER). Models developed based on spectra collected from fillet bone side performed better than those from skin side regardless of color trait, best results were all obtained using 400-2500 nm and closely followed by 400-1100 nm. Prediction results for meat redness a* (R (p) (2) = 0.81; RPD = 2.21; RER = 12.13) and hue angle (R (p) (2) = 0.80; RPD = 2.07; RER = 12.65) were the best and for L*, b*, and chroma (R (p) (2) ae<yen> 0.72; RPD ae<yen> 1.81; RER ae<yen> 8.48) were also well. This work systematically evaluated the influence of different color trait parameters, fillet side spectra, spectral ranges, and pre-treatment methods compared with previous investigation, and results suggest that Vis/NIR spectroscopy can be a feasible method for prediction of color traits of intact broiler breast fillets, especially when spectra are collected from the fillet bone side and range of optimal partial ranges is included.