KR-20260062577-A - Prediction method and system for determining meat quality as a function of its color changes

Abstract

The present invention relates to a method and system for determining meat quality. The meat quality determination method of the present invention can rapidly and accurately determine the quality and freshness of meat, such as beef, using a non-destructive method, and can be utilized in the field of meat distribution and sales.

Inventors

• 도한솔
• 김소민

Assignees

• 이화여자대학교 산학협력단

Dates

Publication Date

20260507

Application Date

20241029

Claims (15)

1. 1) A step of obtaining color information by measuring the color of the target meat; 2) A step of estimating the spoilage indicator of the target meat by using the above color information as input to the first model; 3) A step of estimating the storage period of the target meat by using the above-mentioned estimated spoilage indicator as input to the second model; and 4) A step of determining the quality of the target meat based on the above-mentioned estimated storage period A method for determining the quality of meat, including
2. A method according to claim 1, wherein the color information of the meat is expressed using the CIELAB color space.
3. A method according to claim 1, wherein the color information of the meat includes an a * value in CIELAB units.
4. A method according to claim 1, wherein the spoilage indicator comprises one or more selected from the group consisting of metmyoglobin (Met. Mb) content, pH, peroxide value (PV), and thiobarbituric acid reactive substances (TBARS) values.
5. The method of claim 1, wherein the first model is a model trained with a dataset including color information and spoilage indicator information of meat.
6. A method according to claim 1, wherein the first model comprises one or more algorithms selected from a group consisting of GB (gradient boosting) regression, RF (random forest) regression, KNN (k-nearest neighbor) regression, polynomial regression, SVR (support vector regression), and LASSO (least absolute shrinkage and selection operator) regression.
7. The method of claim 1, wherein the second model is a model trained with a dataset including information on the spoilage indicator of meat and information on the storage period.
8. The method of claim 1, wherein the second model comprises one or more algorithms selected from the group consisting of DT (decision tree), KNN (k-nearest neighbor), SVM (support vector machine), and MLP (multilayer perceptron) neural networks.
9. The method of claim 1, wherein the method further comprises the step of 5-1) determining that the target meat is of fresh quality if the estimated storage period is 0 to 168 hours.
10. The method of claim 1, wherein the method further comprises the step of 5-2) determining that the target meat is of average quality if the estimated storage period is 168 hours to 240 hours.
11. The method of claim 1, wherein the method further comprises the step of 5-3) determining that the target meat is of a quality at the initial stage of spoilage if the estimated storage period is 216 hours or later.
12. A method according to claim 1, wherein the meat is derived from one selected from the group consisting of pig, cattle, sheep, and chicken.
13. In a computer program for determining the quality of meat recorded on a computer-readable recording medium to execute a method for determining the quality of meat combined with a computer system, the program A first model that takes color information of the target meat as an input value and predicts spoilage indicator information of the target meat based on the color information as an output value; and It includes a second model that takes a spoilage index of the target meat as an input value and a storage period of the target meat predicted based on the said spoilage index as an output value, and A computer program capable of determining the quality of the target meat based on the above-mentioned storage period of the meat.
14. In a system for determining the quality of meat implemented by a computer, It includes at least one processor implemented to execute computer-readable instructions, and A meat quality determination system comprising: at least one processor including predicting spoilage indicator information of the target meat by using color information of the target meat as input to a first model; and predicting storage period information of the target meat by using the predicted spoilage indicator information of the target meat as input to a second model.
15. In computer devices, Includes a processor, The above processor is, A computer device that predicts spoilage indicator information of target meat by using color information of target meat as input to a first model, predicts storage period information of target meat by using the predicted spoilage indicator information of target meat as input to a second model, and determines the quality of meat based on the predicted storage period information.

Description

Prediction method and system for determining meat quality as a function of its color changes The present invention relates to a method and system for determining meat quality. Recently, beef consumption has been continuously increasing, and as consumption rises, maintaining beef quality is becoming increasingly important for both the beef processing industry and consumers. Therefore, monitoring and predicting freshness is essential for beef quality control and is effective in preventing diseases that may be caused by beef. Previously, sophisticated techniques using instruments such as Kjeldahl distillation, titration, and spectrophotometry were used to measure the freshness and quality of beef. However, these techniques are destructive methods that incur relatively high costs, require a long duration, and necessitate expert supervision, making them unsuitable for real-time analysis of beef freshness. Therefore, there is a need to develop a technology that can analyze the freshness and quality of beef quickly and simply without being destructive. Meanwhile, artificial intelligence (AI) technology has been advancing recently, and machine learning (ML), a type of AI, can be applied to food science and nutrition, primarily to increase productivity and efficiency in the food processing industry. Accordingly, the inventors developed a method and system capable of determining the quality and freshness of meat, such as beef, and completed the present invention by confirming that the quality and freshness of meat can be predicted accurately and quickly through a machine learning-based model. Figure 1 is a diagram showing the color change of beef according to the storage period. Figure 2 is a diagram showing the change in TVB-N content according to the storage period of beef. The blue dotted line indicates 'freshness', and the red dotted line indicates 'early spoilage stage'. Figure 3 is a diagram showing the changes in the a * value, metmyoglobin content, pH, PV, and TBARS of the CIELAB color space according to the storage period of beef. Figure 4 is a diagram showing the results of the principal component analysis (PCA) of beef spoilage indicators. Figure 5 is a diagram showing the results of the error matrix analysis of the classification model used. Figure 6 is a figure showing the accuracy, precision, recall, and F1-score calculated based on the error matrix. Figure 7 is a diagram showing the correlation between the a * value of the CIELAB color space and the decay index. Figure 8 is a scatter plot showing the predicted and measured values of the beef spoilage index according to the a * value of the CIELAB color space in the GB (gradient boosting) regression model. Figure 9 is a scatter plot showing the predicted and measured values of the beef spoilage index according to the a * value of the CIELAB color space in the RF (random forest) regression model. Figure 10 is a scatter plot showing the predicted and measured values of the beef spoilage index according to the a * value of the CIELAB color space in a polynomial regression model. Figure 11 is a scatter plot showing the predicted and measured values of the beef spoilage index according to the a * value in the CIELAB color space in the KNN (k-nearest neighbor) regression model. Figure 12 is a scatter plot showing the predicted and measured values of the beef spoilage index according to the a * value in the CIELAB color space in the SVR (support vector regression) model. Figure 13 is a scatter plot of the predicted and measured values of the beef spoilage index according to the a * value in the CIELAB color space in the LASSO (least absolute shrinkage and selection operator) regression model. FIG. 14 is a flowchart illustrating an example of a method for determining the quality of meat according to the present invention. FIG. 15 is a schematic diagram showing an example of a meat quality determination system or model of the present invention. The following examples will be explained in more detail. However, these examples are for illustrative purposes only and the scope of the present invention is not limited to these examples. Example 1: Sample Preparation The beef used in this invention was purchased from a local market in Seoul, Korea, and specifically the upper body (Musculus semimembranosus) was used. For the experiment, beef was cut into chunks of approximately 30g, wrapped in polyethylene plastic bags, and stored in a refrigerator at 4°C to prepare samples. The samples were stored for 11 days, and the exact amount required for each subsequent experiment was measured again. The samples were observed periodically at 2-day intervals to check for color changes (Fig. 1). Example 2: Color verification of beef samples The color of the beef was verified using CIELAB. CIELAB color parameters were measured using a non-contact colorimeter (AEROS model, HunterLab, Reston, VA, USA) equipped with D65 illumination and a 10° observation angle. Measurements were perform