Classifier design for computer grading systems for food processing

In recent years there has been an increase in the application of computer grading systems for automatic food processing. Using intelligent sensing and consistent classification rules, computer grading systems can provide an objective solution for a higher level of quality control for food products. Two types of grading are defined: hard grading and fuzzy grading. In this paper, specific guidelines are presented for classifier design for application in quality assessment in food processing. A generalized fuzzy classifier, proposed by Bezdek, et al. (1986), is modified to include optimization for the full benefit of cost saving in food processing. Examples are shown for computer grading of fish product (herring roe) using both hard and fuzzy partition rules