Correlating shrinkage with yield, water content and texture of pork ham by computer vision

An automatic method for shrinkage evaluation of pork ham was developed using computer vision. First, a sequence of image-processing algorithms was developed to estimate the average diameter, short axis, long axis, perimeter, volume and surface area before and after cooking and cooling. This sequence consisted of three steps, i.e., shape extraction, protrusion deletion and measurement. Based on the estimated shape characteristics, three kinds of shrinkage were evaluated as the percentage change before and after a process, i.e., shrinkages caused by the cooking process, cooling process and total shrinkage during the entire cooking and cooling processes. Then the cooking shrinkage was related to cooking loss; the cooling shrinkage to cooling loss and the total shrinkage to yield, water content and texture. It was found that among the three shrinkages, the cooking shrinkage in volume was the highest with up to 9.36%, and was significantly correlated with cooking loss (r=0.91). The total shrinkage was highly negatively correlated with water content, and had positive correlations with the texture attributes. However, no significant relationships were found between cooling shrinkage and cooling loss, and between total shrinkage and yield.