Application of texture image analysis for the classification of bovine meat

Texture analysis has been used to classify photographic images of meat slices. Among the multiple muscular tissue characteristics that influence meat quality, the connective tissue content and spatial distribution, which define the grain of meat, are of great importance because they are directly related to its tenderness. Connective tissue contains two important components, fat and collagen, which are variable with muscles, breed and also with age. These components are clearly visible on photographic images. Fat and collagen are particularly emphasised by ultraviolet light. The meat slices analysed came from 26 animals raised at INRA of Theix by the LCMH Laboratory. Three different muscles were selected and cut off from carcasses of animals of different breeds and of different ages. The biological factors (muscle type, age and breed) directly influence the structure and composition of the muscle samples. The image analysis led to a representation of each meat sample with a 58 features vector. Classification experiments were performed to identify the samples according to the three variation factors. This study shows the potential of image analysis for meat sample recognition. The correlation of the textural features with chemical and mechanical parameters measured on the meat samples was also examined. Regression experiments showed that textural features have potential to indicate meat characteristics.