In beef, the marbling flecks are determinant regarding to flavor and tenderness. Beef cuts with higher levels of marbling are likely to be tenderer, flavor full and juicier than cuts with lower marbling levels [1]. Therefore, the USDA (United States Department of Agriculture) quality grade standards are based on subjective evaluation of the relative degree of visible intramuscular fat [2]. There are 10 official marbling classes: Devoid (D) practically devoid (PD), traces (TR), slight (SL), small (SM), modest (MT), moderate (MD) slightly abundant (SA), moderately abundant (MA) and abundant (AB). So, there is a growing interest in developing methods and techniques for beef quality classification using digital images. This paper presents an automatic methodology, based on image processing techniques, to identify and to locate the beef in the image and to calculate the marbling measures, in order to evaluate the beef quality. The tests realized showed that it is possible to use image analysis in color photographs of beef steaks to automatically extract the marbling measures and to identify the beef quality. However, to develop more accurate algorithms, larger training and testing sets must be used.
CITATION STYLE
Caridade, C. M. R., Pereira, C. D., Pires, A. F., Marnotes, N. G., & Viegas, J. F. (2019). Automatic Extraction of Marbling Measures Using Image Analysis, for Evaluation of Beef Quality. In Lecture Notes in Computational Vision and Biomechanics (Vol. 34, pp. 437–446). Springer Netherlands. https://doi.org/10.1007/978-3-030-32040-9_44
Mendeley helps you to discover research relevant for your work.