Detection and classification of meat freshness using an optimized deep learning method

Accurate assessment of meat freshness is critical for ensuring food safety, reducing waste, and maintaining quality control in the food industry. Efficient classification of meat into categories such as fresh, half-fresh, and spoiled is essential for producers, retailers, and consumers alike. This paper presents a deep learning-based method for meat freshness classification, comprising three main stages: (1) feature extraction using the Visual Geometry Group 19 (VGG19) convolutional neural network; (2) feature selection using an Improved Artificial Protozoa Optimizer (IAPO) enhanced by Particle Swarm Optimization (PSO); and (3) classification based on the optimized feature set. The proposed method was benchmarked against five established optimization techniques. Experimental results demonstrate that the method achieves an accuracy of 98.51%, sensitivity of 98.54%, and specificity of 99.24%. These findings indicate that the method may be suitable for meat freshness classification in food safety monitoring systems.