Real-time monitoring the color changes of large yellow croaker (Larimichthys crocea) fillets based on hyperspectral imaging empowered with artificial intelligence

Rapid and non-destructive microbial quality prediction of fresh pork stored under modified atmospheres by using selected-ion flow-tube mass spectrometry and machine learning

Volatile organic compounds (VOCs) indicative of pork microbial spoilage can be quantified rapidly at trace levels using selected-ion flow-tube mass spectrometry (SIFT-MS). Packaging atmosphere is one of the factors influencing VOC production patterns during storage. On this basis, machine learning would help to process complex volatolomic data and predict pork microbial quality efficiently. This study focused on (1) investigating model generalizability based on different nested cross-validation settings, and (2) comparing the predictive power and feature importance of nine algorithms, including Artificial Neural Network (ANN), k-Nearest Neighbors, Support Vector Regression, Decision Tree, Partial Least Squares Regression, and four ensemble learning models. The datasets used contain 37 VOCs' concentrations (input) and total plate counts (TPC, output) of 350 pork samples with different storage times, including 225 pork loin samples stored under three high-O2 and three low-O2 conditions, and 125 commercially packaged products. An appropriate choice of cross-validation strategies resulted in trustworthy and relevant predictions. When trained on all possible selections of two high-O2 and two low-O2 conditions, ANNs produced satisfactory TPC predictions of unseen test scenarios (one high-O2 condition, one low-O2 condition, and the commercial products). ANN-based bagging outperformed other employed models, when TPC exceeded ca. 6 log CFU/g. VOCs including benzaldehyde, 3-methyl-1-butanol, ethanol and methyl mercaptan were identified with high feature importance. This elaborated case study illustrates great prospects of real-time detection techniques and machine learning in meat quality prediction. Further investigations on handling low VOC levels would enhance the model performance and decision making in commercial meat quality control.

Off-flavor profiling of cultured salmonids using hyperspectral imaging combined with machine learning

Off-flavors can have significant impacts on the quality of salmonid products. This study investigated the possibility of comprehensive off-flavor profiling considering both olfactory and taste sensory perspectives by combining near-infrared hyperspectral imaging (NIR-HSI) and machine/deep learning. Four feature extraction algorithms were employed for the extraction and interpretation of spectral fingerprint information regarding off-flavor-related compounds. Classification models, including the partial least squares discriminant analysis, least-squares support vector machine, extreme learning machine, and one-dimensional convolutional neural network (1DCNN) were constructed using the full wavelengths and selected spectral features for the identification of off-flavor salmonids. The 1DCNN achieved the highest discrimination accuracy with full and selected wavelengths (i.e., 91.11 and 86.39 %, respectively). Furthermore, the prediction and visualization of off-flavor-related compounds were achieved with acceptable performances (R² > 0.6) for practical applications. These results indicate the potential of NIR-HSI for the off-flavor profiling of salmonid muscle samples for producers and researchers.

Nondestructive detection of nutritional parameters of pork based on NIR hyperspectral imaging technique

Nondestructive detection of the nutritional parameters of pork is of great importance. This study aimed to investigate the feasibility of applying hyperspectral image technology to detect the nutrient content and distribution of pork nondestructively. Hyperspectral cubes of 100 pork samples were collected using a line-scan hyperspectral system, the effects of different preprocessing methods on the modeling effects were compared and analyzed, the feature wavelengths of fat and protein were extracted, and the full-wavelength model was optimized using the regressor chains (RC) algorithm. Finally, pork's fat, protein, and energy value distributions were visualized using the best prediction model. The results showed that standard normal variate was more effective than other preprocessing methods, the feature wavelengths extracted by the competitive adaptive reweighted sampling algorithm had better prediction performance, and the protein model prediction performance was optimized after using the RC algorithm. The best prediction models were developed, with the correlation coefficient of prediction (R_P) = 0.929, the root mean square error in prediction (RMSEP) = 0.699% and residual prediction deviation (RPD) = 2.669 for fat, and R_P = 0.934, RMSEP = 0.603% and RPD = 2.586 for protein. The pseudo-color maps were helpful for the analysis of nutrient distribution in pork. Hyperspectral image technology can be a fast, nondestructive, and accurate tool for quantifying the composition and assessing the distribution of nutrients in pork.

Real-Time Monitoring of the Quality Changes in Shrimp (Penaeus vannamei) with Hyperspectral Imaging Technology during Hot Air Drying

Hot air drying is the most common processing method to extend shrimp’s shelf life. Real-time monitoring of moisture content, color, and texture during the drying process is important to ensure product quality. In this study, hyperspectral imaging technology was employed to acquire images of 104 shrimp samples at different drying levels. The water distribution and migration were monitored by low field magnetic resonance and the correlation between water distribution and other quality indicators were determined by Pearson correlation analysis. Then, spectra were extracted and competitive adaptive reweighting sampling was used to optimize characteristic variables. The grey-scale co-occurrence matrix and color moments were used to extract the textural and color information from the images. Subsequently, partial least squares regression and least squares support vector machine (LSSVM) models were established based on full-band spectra, characteristic spectra, image information, and fused information. For moisture, the LSSVM model based on full-band spectra performed the best, with residual predictive deviation (RPD) of 2.814. For L*, a*, b*, hardness, and elasticity, the optimal models were established by LSSVM based on fused information, with RPD of 3.292, 2.753, 3.211, 2.807, and 2.842. The study provided an in situ and real-time alternative to monitor quality changes of dried shrimps.

Recent advances in muscle food safety evaluation: Hyperspectral imaging analyses and applications

As there is growing interest in process control for quality and safety in the meat industry, by integrating spectroscopy and imaging technologies into one system, hyperspectral imaging, or chemical or spectroscopic imaging has become an alternative analytical technique that can provide the spatial distribution of spectrum for fast and nondestructive detection of meat safety. This review addresses the configuration of the hyperspectral imaging system and safety indicators of muscle foods involving biological, chemical, and physical attributes and other associated hazards or poisons, which could cause safety problems. The emphasis focuses on applications of hyperspectral imaging techniques in the safety evaluation of muscle foods, including pork, beef, lamb, chicken, fish and other meat products. Although HSI can provide the spatial distribution of spectrum, characterized by overtones and combinations of the C-H, N-H, and O-H groups using different combinations of a light source, imaging spectrograph and camera, there still needs improvement to overcome the disadvantages of HSI technology for further applications at the industrial level.