The study demonstrated the potentials for automatic and objective
determination of pork marbling scores. Prediction results of
the three models showed that the prediction ability of the simple
linear model LR_B was comparable with the multiple linear model
MLR_RGB. This shows the efficiency of the marbling extraction
procedure, alleviating the contrast problem brought about by PSE
and PFN samples when subjective marbling assessment is made
on these pale and reflective pork samples.
The novelty of this paper is to propose an automatic and objective
evaluation method for pork marbling score assessment in the
view of pattern recognition. Marbling scores from standard charts
have been used in the study. Considering the relationship between
marbling scores and intramuscular fat (IMF) content, future work
will focus on studying the potential of predicting IMF content of
pork using hyperspectral imaging and pattern recognition
technique.