Analysis of beef quality according to color changes using computer vision and white-box machine learning techniques

ConvColor DL: Concatenated convolutional and handcrafted color features fusion for beef quality identification

Beef is an important food product in human nutrition. The evaluation of the quality and safety of this food product is a matter that needs attention. Non-destructive determination of beef quality by image processing methods shows great potential for food safety, as it helps prevent wastage. Traditionally, beef quality determination by image processing methods has been based on handcrafted color features. It is, however, difficult to determine meat quality based on the color space model alone. This study introduces an effective beef quality classification approach by concatenating learning-based global and handcrafted color features. According to experimental results, the convVGG16 + HLS + HSV + RGB + Bi-LSTM model achieved high performance values. This model's accuracy, precision, recall, F1-score, AUC, Jaccard index, and MCC values were 0.989, 0.990, 0.989, 0.990, 0.992, 0.979, and 0.983, respectively.

AI-based R&D for frozen and thawed meat: Research progress and future prospects

Frozen and thawed meat plays an important role in stabilizing the meat supply chain and extending the shelf life of meat. However, traditional methods of research and development (R&D) struggle to meet rising demands for quality, nutritional value, innovation, safety, production efficiency, and sustainability. Frozen and thawed meat faces specific challenges, including quality degradation during thawing. Artificial intelligence (AI) has emerged as a promising solution to tackle these challenges in R&D of frozen and thawed meat. AI's capabilities in perception, judgment, and execution demonstrate significant potential in problem-solving and task execution. This review outlines the architecture of applying AI technology to the R&D of frozen and thawed meat, aiming to make AI better implement and deliver solutions. In comparison to traditional R&D methods, the current research progress and promising application prospects of AI in this field are comprehensively summarized, focusing on its role in addressing key challenges such as rapid optimization of thawing process. AI has already demonstrated success in areas such as product development, production optimization, risk management, and quality control for frozen and thawed meat. In the future, AI-based R&D for frozen and thawed meat will also play an important role in promoting personalization, intelligent production, and sustainable development. However, challenges remain, including the need for high-quality data, complex implementation, volatile processes, and environmental considerations. To realize the full potential of AI that can be integrated into R&D of frozen and thawed meat, further research is needed to develop more robust and reliable AI solutions, such as general AI, explainable AI, and green AI.

Measuring water holding capacity in poultry meat

In the current scientific literature, one can find >100 different methods to evaluate water-holding capacity in fresh and cooked meat. The main concepts are based on removing some of the water by either gravity, application of pressure (e.g., centrifugal force), and heating while measuring water exudate to predict the water holding capacity (WHC) during storage, processing, cooking, and/or distribution. More sophisticated methods include nuclear magnetic resonance (NMR) in which the relaxation of water molecules within a meat protein/gel system is measured to predict how the water (75% in lean meat) will behave during processing. Overall, the number of tests reported is also so high because there are quite big variations in test conditions (e.g., 750-30,000 g for centrifugal testing). The aim of this article (outcome of a symposium on methods for poultry meat characterization) is to help the reader navigate through the different setups and suggest standardized testing based on scientific principles. The recommended WHC test is the application of low centrifugal force (750 g so sample is not permanently deformed) to a protein gel, while the sample is placed on a screen platform to avoid reabsorbing the liquid separating during the slowing down of the centrifuge. It is also recognized that some meat samples (e.g., high in fat) might require a different g-force, so it is recommended to employ both the conditions mentioned above and the lab-specific conditions. Our overall goal should always be to increase uniformity in test procedures, which will enhance our capabilities to compare results among research groups.

Research Progress on Quality Detection of Livestock and Poultry Meat Based on Machine Vision, Hyperspectral and Multi-Source Information Fusion Technologies

Presently, the traditional methods employed for detecting livestock and poultry meat predominantly involve sensory evaluation conducted by humans, chemical index detection, and microbial detection. While these methods demonstrate commendable accuracy in detection, their application becomes more challenging when applied to large-scale production by enterprises. Compared with traditional detection methods, machine vision and hyperspectral technology can realize real-time online detection of large throughput because of their advantages of high efficiency, accuracy, and non-contact measurement, so they have been widely concerned by researchers. Based on this, in order to further enhance the accuracy of online quality detection for livestock and poultry meat, this article presents a comprehensive overview of methods based on machine vision, hyperspectral, and multi-sensor information fusion technologies. This review encompasses an examination of the current research status and the latest advancements in these methodologies while also deliberating on potential future development trends. The ultimate objective is to provide pertinent information and serve as a valuable research resource for the non-destructive online quality detection of livestock and poultry meat.

Images dataset of beef meat samples with different shelf life

Three different cuts of meat samples: inside skirt, knuckles, and sirloin were picture captioned on the first and fifth day after purchase. From each type of meat cut, ten pictures were taken at the beginning and the end of the studied shelf life, obtaining 60 different images. The images were taken under control variables in a black acrylic cabin. In addition to the original images, we proportionate another set of 60 processed images. The latter were obtained after color calibration and meat segmentation. All these images could be used for future experiments where the color in meat should be analyzed.