Predictive efficiency of distinct color image segmentation methods for measuring intramuscular fat in beef

Ciência Rural

Renius Mello¹  Fabiano Nunes Vaz¹  Paulo Santana Pacheco¹  Leonir Luiz Pascoal¹  Rosa Cristina Prestes¹  Patrícia Barcellos Costa¹  Djenifer Kirch Kipper¹ 

关键词: tissue characterization;    meat industry;    marbling;    image processing;    meat quality.;    caracterização tecidual;    indústria cárnea;    marmoreio;    processamento de imagens;    qualidade da carne.;   

DOI  :  10.1590/0103-8478cr20141617

来源: SciELO

【 摘 要 】

Intramuscular fat (IMF) influences important quality characteristics of meat, such as flavor, juiciness, palatability, odor and tenderness. Thus, the objective of this study was to apply the following image processing techniques to quantify the IMF in beef: palette; sampling, interval of coordinates; black and white threshold; and discriminant function of colors. Thirty-five samples of beef, with a wide range of IMF, were used. Color images were taken of the meat samples from different muscles, with variability in the IMF content. The IMF of a thin cross-section meat was determined by chemical lipid extraction and was predicted by image analysis. The chemical method was compared with the image analysis. The segmentation procedures were validated by the adjustment of a linear regression equation to the series of values that were observed and predicted, as well as the regression parameters evaluated by the F-test. The predictive power of these approaches was also compared by residual analysis and by the decomposition of the mean square deviations. The results showed that the discriminant function was the best color segmentation method to measure intramuscular fat via digital images, but required adjustments in the prediction pattern.

【 授权许可】

CC BY   
 All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License