Machine Learning Approaches to Understanding Meat Color Changes via Digital Imaging
نویسندگان
1 Chemistry Department, Sharif University of Technology, Tehran, Iran
2 Halal Research Center of IRI, Food and Drug Administration, Ministry of Health and Medical Education, Tehran, Iran.
3 Department of Food and Drug Control, School of Pharmacy, Tehran University of Medical Sciences, Tehran, Iran
4 Halal Research Center of IRI, Food and Drug Administration, Ministry of Health and Medical Education, Tehran, Iran
5 Halal Research Center of IRI, Food and Drug Administration, Ministry of Health and Medical Education, Tehran, Iran.
6 Cosmetic Products Research Center, Iran Food and Drug Administration, MOHE, Tehran, Iran, Future Studies Group, the Academy of Medical Sciences of The I.R. Iran, Tehran, Iran
doi
10.22036/abcr.2025.503639.2277چکیده
Ensuring meat quality is critical for consumer satisfaction and food safety, but monitoring subtle quality changes remains a challenge. This study presents a smartphone-based imaging approach combined with advanced data analysis techniques to evaluate color changes in red meat under different storage conditions. Meat samples stored in a refrigerator (4°C) and a freezer (-19°C) were analyzed over three weeks using RGB and HSV color spaces. Principal Component Analysis (PCA) revealed patterns of color change, while ANOVA-Simultaneous Component Analysis (ASCA) identified significant effects of storage time and temperature on meat color, with the HSV color space showing greater sensitivity. Partial Least Squares Discriminant Analysis (PLS-DA) successfully classified chilled and frozen samples after temperature equilibration, with the soft classification method demonstrating robust performance. These results highlight the potential of integrating accessible imaging tools and machine learning techniques for objective and efficient meat quality assessment, providing a scalable solution for the food industry.
