Method and apparatus for determining grade of meat by using artificial intelligence model

Abstract

One embodiment of the present disclosure provides a method for determining the grade of meat by using an artificial intelligence model. The method comprises the steps of: generating a training data set including a training meat image of at least one from among Korean pork, Korean beef, and beef and grade information matched with the training meat image; and generating an artificial intelligence model trained to determine, on the basis of the training data set, the grade of meat included in a meat image that is input. The present invention can provide a consistent and highly accurate meat grade evaluation by means of artificial intelligence having been trained with meat images classified by grade. In addition, the present invention can: be performed by artificial intelligence, thereby reducing costs and improving evaluation speed; provide a meat grade evaluation suitable for the grade of the corresponding country by evaluating, by country, with the same criteria; and provide a basis on which an item has played a decisive role in evaluating the grade by clarifying meat grade evaluation criteria.

Description

Method and device for judging meat grade using an artificial intelligence model

[0001] The present invention relates to a method and device for judging meat grade using an artificial intelligence model, and more specifically, to a technology for analyzing a meat image using an artificial intelligence model and judging the grade of the meat included in the image.

[0002] Meat consumption continues to rise globally. As of 2021, chicken ranked first in global meat consumption, followed by pork and beef.

[0003] In 2021, pork consumption reached 32.3 kg, while domestic beef consumption reached 12.4 kg per capita. Today's consumers have access to information about meat through various media outlets, and they use this information to purchase higher-quality meat. This is because there are significant differences in taste and price depending on the grade of meat.

[0004] However, meat grading is difficult for anyone but experts, and opinions can vary widely. This is because key criteria for meat grading—marbling, color, texture, and palatability—vary among experts and across countries.

[0005] The lack of objective data to support the evaluation results is the biggest reason for the decline in reliability of meat grading.

[0006] Accordingly, there is an increasing need for a meat grading method and device using artificial intelligence that analyzes a single meat photo and determines the grade based on data held by artificial intelligence that has learned a number of meat photos by grade.

[0007] The present invention is intended to solve the problems of the prior art described above, and relates to a method and device for judging meat grade using artificial intelligence, in which pre-learned artificial intelligence analyzes meat included in an image and judges the grade.

[0008] However, the technical task that this embodiment seeks to achieve is not limited to the technical task described above, and other technical tasks may exist.

[0009] As a technical means for achieving the above-described technical task, an embodiment according to the first aspect of the present disclosure provides a method for judging meat grade using an artificial intelligence model. The method comprises the steps of: generating a learning data set including learning meat images for at least one of Korean pork, Korean beef, and beef cattle, and grade information matching the learning meat images; and generating an artificial intelligence model trained to judge the grade of meat included in an input meat image based on the learning data set.

[0010] In addition, an embodiment according to a second aspect of the present disclosure provides a meat grading device. The device includes a communication module, at least one processor, and a memory electrically connected to the processor and storing at least one code executed by the processor, wherein the memory stores a code that, when executed through the processor, causes the processor to generate a learning data set including learning meat images for at least one of Korean pork, Korean beef, and beef cattle and grade information matching the learning meat images, and generate an artificial intelligence model trained to determine the grade of meat included in an input meat image based on the learning data set.

[0011] The present invention provides consistent and highly accurate meat grading using artificial intelligence trained on graded meat images. Furthermore, the AI-based approach reduces costs and improves evaluation speed.

[0012] Additionally, by evaluating the same standards for each country, we can provide a meat grading evaluation that matches the grade of that country.

[0013] Additionally, by clarifying the criteria for meat grading, it can provide evidence on which items played a decisive role in evaluating the grade.

[0014] FIG. 1 is a drawing showing a meat grading device and a terminal connected to the device in communication with the device according to one embodiment of the present invention.

[0015] Figure 2 is a detailed drawing of the meat grading device in Figure 1.

[0016] FIG. 3 is a diagram illustrating grade prediction according to one embodiment of the present invention.

[0017] Figure 4 is a diagram illustrating a process for obtaining a sensory evaluation according to one embodiment of the present invention.

[0018] Figure 5 is a diagram showing a grade prediction result using an artificial intelligence model according to one embodiment of the present invention.

[0019] FIG. 6 is a diagram showing a grade prediction result through an attention map according to one embodiment of the present invention.

[0020] Figure 7 is a flowchart illustrating the sequence of a meat grading method according to another embodiment of the present invention.

[0021] Hereinafter, the present disclosure will be described in detail with reference to the attached drawings. However, the present disclosure may be implemented in various different forms and is not limited to the embodiments described herein. In addition, the attached drawings are only intended to facilitate understanding of the embodiments disclosed in the present specification, and the technical concepts disclosed in the present specification are not limited by the attached drawings. All terms, including technical and scientific terms, used herein should be interpreted as having meanings generally understood by a person of ordinary skill in the technical field to which the present disclosure pertains. Terms defined in the dictionary should be interpreted as having additional meanings consistent with the relevant technical literature and the contents of the present disclosure, and shall not be interpreted in an extremely ideal or restrictive sense unless otherwise defined.

[0022] In order to clearly explain the present invention in the drawings, parts irrelevant to the description have been omitted, and the size, shape, and appearance of each component shown in the drawings may be modified in various ways. Identical / similar parts throughout the specification are given identical / similar drawing reference numerals.

[0023] Throughout the specification, when a part is said to be "connected (connected, in contact with, or coupled)" to another part, this includes not only cases where it is "directly connected (connected, in contact with, or coupled)" but also cases where it is "indirectly connected (connected, in contact with, or coupled)" with another member in between. Furthermore, when a part is said to "include (have or provide)" a certain component, this does not mean that it excludes other components, but rather that it may "include (have or provide)" other components, unless otherwise specifically stated.

[0024] In this specification, the term 'unit' includes a unit realized by hardware, a unit realized by software, and a unit realized using both. In addition, one unit may be realized by using two or more pieces of hardware, and two or more units may be realized by one piece of hardware. Meanwhile, the '~ unit' is not limited to software or hardware, and the '~ unit' may be configured to be in an addressable storage medium or may be configured to reproduce one or more processors. Therefore, as an example, the '~ unit' includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functionality provided within the components and '~sub-units' may be combined into a smaller number of components and '~sub-units' or further separated into additional components and '~sub-units'. Furthermore, the components and '~sub-units' may be implemented to activate one or more CPUs within the device or secure multimedia card.

[0025] The suffixes "module" and "part" used in the following description for components are assigned or used interchangeably solely for the convenience of writing the specification, and do not in themselves have distinct meanings or roles. Furthermore, in describing the embodiments disclosed herein, detailed descriptions of related known technologies have been omitted if they are deemed to obscure the gist of the embodiments disclosed herein.

[0026] As used herein, ordinal terms such as "first," "second," etc., are used solely to distinguish one component from another and do not limit the order or relationship of the components. For example, the first component of the present disclosure may be referred to as the "second component," and similarly, the second component may also be referred to as the "first component." As used herein, singular forms should be construed to include plural forms, unless explicitly stated otherwise.

[0027] The "user terminal" mentioned below may be implemented as a computer or portable terminal that can access a server or other terminal via a network. Here, the computer may include, for example, a notebook, desktop, laptop, VR HMD (e.g., HTC VIVE, Oculus Rift, GearVR, DayDream, PSVR, etc.) equipped with a web browser. Here, the VR HMD includes all of the stand-alone models implemented independently for PC (e.g., HTC VIVE, Oculus Rift, FOVE, Deepon, etc.), mobile (e.g., GearVR, DayDream, Storm Magic, Google Cardboard, etc.), and console (PSVR) (e.g., Deepon, PICO, etc.). A portable terminal is, for example, a wireless communication device that ensures portability and mobility, and may include not only a smart phone, a tablet PC, and a wearable device, but also various devices equipped with communication modules such as Bluetooth (BLE, Bluetooth Low Energy), NFC, RFID, ultrasonic, infrared, WiFi, and LiFi. In addition, a "network" refers to a connection structure that enables information exchange between each node, such as terminals and servers, and includes a local area network (LAN), a wide area network (WAN), the Internet (WWW: World Wide Web), wired and wireless data communication networks, telephone networks, and wired and wireless television communication networks.Examples of wireless data communication networks include, but are not limited to, 3G, 4G, 5G, 3GPP (3rd Generation Partnership Project), LTE (Long Term Evolution), WIMAX (World Interoperability for Microwave Access), Wi-Fi, Bluetooth, infrared, ultrasonic, visible light communication (VLC), and LiFi.

[0028] FIG. 1 is a drawing showing a meat grading device and a terminal connected to the device in communication with the device according to one embodiment of the present invention.

[0029] Referring to FIG. 1, a meat grading device (100) and a terminal (200) can be connected to each other through a communication network. Hereinafter, the meat grading device (100) may be referred to as a server. The server may be implemented as a cloud computing server such as SaaS (Software as a Service), PaaS (Platform as a Service), or IaaS (Infrastructure as a Service). In addition, the server may be implemented in the form of a private cloud, a public cloud, or a hybrid cloud system, but the scope of the present invention is not limited thereto. In addition, the meat grading device (100) may be driven using an artificial intelligence model.

[0030] The meat grading device (100) generates a learning data set including at least one learning meat image and grade information matching the learning meat image.

[0031] The meat grading device (100) creates an artificial intelligence model trained to judge the grade of meat included in an input meat image based on a learning data set.

[0032] The terminal (200) can transmit an image to the meat grading device (100). Then, the terminal can receive the meat grade from the meat grading device (100) along with visualization data.

[0033] Figure 2 is a drawing showing in detail the meat grading device (100) in Figure 1.

[0034] Referring to FIG. 2, the meat grading device (100) may include a communication module (110), a processor (120), and a memory (130).

[0035] The communication module (110) may include a device including hardware and software necessary to transmit and receive signals such as control signals or data signals through wired or wireless connections with other network devices.

[0036] The communication module (110) can receive meat images from the terminal (200). In addition, the communication module (110) can provide meat grades to the terminal (200) as visual data including images, graphs, and tables. However, this is not limited thereto, and, if necessary, the communication module (110) can receive images or videos from external devices and databases other than the terminal (200).

[0037] The processor (120) may include various types of devices that control and process data. The processor (120) may refer to a data processing device built into hardware that has a physically structured circuit to perform functions expressed in code or instructions included in a program.

[0038] In one example, the processor (120) may be implemented in the form of a microprocessor, a central processing unit (CPU), a processor core, a multiprocessor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), etc., but the scope of the present invention is not limited thereto.

[0039] The processor (120) performs operations according to the code stored in the memory (130).

[0040] The memory (130) can store at least one of information and data input to the communication module (110), information and data required for functions performed by the processor (120), and data generated according to the execution of the processor (120).

[0041] Memory (130) should be interpreted as a general term for a non-volatile storage device that maintains stored information even when no power is supplied, and a volatile storage device that requires power to maintain stored information. Memory (130) may include magnetic storage media or flash storage media in addition to a volatile storage device that requires power to maintain stored information, but the scope of the present invention is not limited thereto.

[0042] The memory (130) is electrically connected to the processor (120) and stores at least one code to be executed by the processor (120). The memory (130) stores code that, when executed by the processor (120), causes the processor (120) to perform the following functions and procedures.

[0043] The memory (130) stores code that causes a learning data set to be generated, including at least one training meat image and grade information matching the training meat image. For example, the grade information may be set based on criteria for determining a plurality of pre-established national beef or pork grades.

[0044] The memory (130) stores a code that causes an artificial intelligence model trained to determine the grade of meat included in an input meat image based on a learning data set.

[0045] The memory (130) receives a meat image to be graded, and stores code that causes an artificial intelligence model to determine the grade of the meat image. For example, the grade of the meat included in the meat image may include at least one grade among multiple national grades.

[0046] When the memory (130) receives a request for a grade judgment of a specific country from a terminal (200) connected to communication, a code is stored that causes the meat included in the meat image to be judged based on a standard corresponding to the grade of the specific country.

[0047] The memory (130) stores code that causes data to be generated that visualizes a region of the meat image as the basis for the determined meat grade. For example, the region may be a region of meat that contains a certain amount or more of the protein contained in the meat image.

[0048] The memory (130) stores a code that causes the terminal (200) connected to the processor (120) to provide an interface including a grade, an evaluation score, a meat image, a QR code, and visualization data.

[0049] FIG. 3 is a diagram illustrating a grade prediction process according to one embodiment of the present invention.

[0050] Referring to FIG. 3, the meat grading device (100) can, upon receiving a meat image from a terminal (200), segment and cut out the portion of the meat to be evaluated, and then resize the image (31). This minimizes information loss. Furthermore, transfer learning can be performed on the resized image using a pre-trained artificial intelligence model. Subsequently, the meat grade can be determined (32) using the artificial intelligence model.

[0051] At this time, the artificial intelligence models used for transfer learning can utilize CNN (Convolutional Neural Network) and ViT (Vision Transformer).

[0052] CNN is a deep learning model that detects image features through convolution operations and uses these to classify input data or extract meaningful information and learn.

[0053] ViT is a deep learning model that uses an attention mechanism for image processing, and is effective in extracting and classifying global information by converting images into tokens.

[0054] More specifically, ViT implements image processing within a Transformer-structured model, leveraging the Transformer's multi-head attention structure rather than using a CNN for vision tasks. This allows ViT to achieve superior performance compared to other CNN-based models and consumes less computational resources during the training process.

[0055] At this time, large values ​​are emphasized based on the Q (Query), K (Key), and V (Value) values, which are the key element representations for learning.

[0056] Learning management for AI models like the above can be accomplished using pre-configured libraries. These libraries support tracking machine learning model experiments and model sharing.

[0057] Figure 4 is a diagram illustrating a process for obtaining a sensory evaluation table according to one embodiment of the present invention.

[0058] Referring to Fig. 4, a meat image can be received from a terminal (200) and a color distribution graph of the corresponding grade can be combined and displayed. At this time, the meat image and the color distribution graph can be arranged to overlap (41) or the meat image and the color distribution graph can be arranged side by side (42). In addition, a sensory evaluation table (44) can be obtained by utilizing an artificial intelligence model (43). The sensory evaluation table (44) can include marbling, meat color, texture, surface juice, and preference.

[0059] FIG. 5 is a diagram visualizing the results of a grade prediction using an artificial intelligence model according to one embodiment of the present invention.

[0060] Referring to Fig. 5, it can be confirmed that the contribution of each grade of meat is visualized using a color distribution graph for each grade. In Fig. 5, a meat image (51) can be combined with a color distribution graph (52) to be expressed as a color distribution image (53). In addition, the target grades of meat are divided into 1++, 1+ grade, 1 grade, 2 grade, and 3 grade, and the results of visualizing which part contributes the most by applying Grad-CAM technology to each grade of the corresponding meat can be confirmed. This means that the closer to red in the meat image, the more the part contributed to the model's prediction. For example, in Fig. 5, the grade of the meat is determined based on the part marked in red in each of the meat images of 1++ grade to 3 grade (54 to 58).

[0061] Grad-CAM (Gradient-weighted Class Activation Mapping) is a technique for visualizing the prediction results of deep learning models. It can highlight the contribution of specific parts of an image to the model's decisions. This allows for a visual understanding of which parts the model primarily utilizes and facilitates interpretation of the model's predictions.

[0062] FIG. 6 is a diagram showing a grade prediction result through an attention map according to one embodiment of the present invention.

[0063] Referring to FIG. 6, this is a technique in which a meat grading device (100) receives a meat image from a terminal (200), creates a square latent vector using an artificial intelligence model, ViT (Vision Transformer) based model, and then resizes it to the same size as the original image to visualize it (61 to 63). For example, as in FIG. 6, the closer it is to the blue color expressed in the meat image, the greater the weight it has.

[0064] The techniques used to visualize the image can utilize Grad-CAM (Gradient-weighted Class Activation Mapping) and Attention Map described in Fig. 5.

[0065] Figure 7 is a flowchart illustrating the sequence of a meat grading method according to another embodiment of the present invention.

[0066] The elevator safety door control method described below can be performed by the meat grading device (100) described above with reference to FIGS. 1 to 6. Therefore, the content of the embodiment of the present disclosure described above with reference to FIGS. 1 to 6 can be equally applied to the embodiment described below, and any content overlapping with the above description will be omitted. The steps described below do not necessarily have to be performed in order, and the order of the steps can be set in various ways, and the steps can be performed almost simultaneously.

[0067] Referring to FIG. 7, the method for judging the grade of meat includes a learning data set generation step (S100), an artificial intelligence model generation step (S200), a step for judging the grade of a meat image (S300), a step for creating visualized data (S400), and a step for providing an interface (S500).

[0068] The learning data set generation step (S100) is a step of generating a learning data set including at least one learning meat image and grade information matching the learning meat image.

[0069] The artificial intelligence model creation step (S200) is a step of creating an artificial intelligence model trained to determine the grade of meat included in an input meat image based on the above learning data set.

[0070] Step (S300) for determining the grade of a meat image is a step of receiving a meat image to be graded and using the artificial intelligence model to determine the grade of the meat image. Alternatively, when a request for grade determination from a specific country is received from a terminal (200) connected to the processor, the meat included in the meat image can be determined and provided based on criteria corresponding to the grade of the specific country.

[0071] The visualized data generation step (S400) is a step of generating data visualizing a region of the meat image as a basis for the judged meat grade.

[0072] The step of providing an interface (S500) is a step of providing an interface including a grade, evaluation score, meat image, QR code, and visualization data to a terminal (200) that is connected to the processor.

[0073] An embodiment of the present invention may also be implemented in the form of a recording medium containing computer-executable instructions, such as program modules, executed by a computer. Computer-readable media may be any available media that can be accessed by a computer, and includes both volatile and nonvolatile media, removable and non-removable media. Furthermore, computer-readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data.

[0074] Although the methods and systems of the present invention have been described with respect to specific embodiments, some or all of their components or operations may be implemented using a computer system having a general-purpose hardware architecture.

[0075] Those skilled in the art will appreciate that the present disclosure can be easily modified into other specific forms based on the above description without changing the technical spirit or essential characteristics of the present disclosure. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. The scope of the present disclosure is indicated by the following claims, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included in the scope of the present disclosure. The scope of the present application is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included in the scope of the present application.

Claims

1. A method for judging meat grades using an artificial intelligence model performed by a processor, a) generating a learning data set including at least one learning meat image and grade information matching the learning meat image; and b) A method for judging the grade of meat using an artificial intelligence model, comprising a step of creating an artificial intelligence model trained to judge the grade of meat included in an input meat image based on the above learning data set.

2. In paragraph 1, c) A method for grading meat using an artificial intelligence model, further comprising the step of receiving a meat image to be graded and using the artificial intelligence model to grade the meat image.

3. In paragraph 1, A method for judging meat grades using an artificial intelligence model, wherein the above grade information is set based on criteria for judging multiple pre-established national beef or pork grades.

4. In paragraph 2, The grade of the meat included in the meat image determined in the above step c) includes at least one grade among multiple national grades, Step c) above, A method for judging meat grades using an artificial intelligence model, wherein, when a request for judging a grade of a specific country is received from a terminal connected to the processor, meat included in the meat image is judged and provided based on criteria corresponding to the grade of the specific country.

5. In paragraph 1, d) A method for judging the grade of meat using an artificial intelligence model, further comprising a step of generating data visualizing a region of the meat image as a basis for the grade of meat judged according to the step c).

6. In paragraph 5, A method for judging meat grade using an artificial intelligence model, wherein the above-mentioned area is a meat area containing a certain portion or more of the protein contained in the meat image.

7. In paragraph 1, e) A method for judging meat grade using an artificial intelligence model, further comprising a step of providing an interface including grades, evaluation scores, meat images, QR codes, and visualization data to a terminal connected to the processor.

8. Communication module; at least one processor; and A memory electrically connected to the processor and storing at least one code to be executed by the processor, The above memory, when executed through the processor, causes the processor to: A meat grading device storing code that causes a learning data set including at least one learning meat image and grade information matching the learning meat image, and to generate an artificial intelligence model trained to judge the grade of meat included in an input meat image based on the learning data set.

9. In paragraph 8, The above memory, when executed through the processor, causes the processor to: A meat grading device that receives a meat image to be graded, grades the meat image using the artificial intelligence model, generates data visualizing a region of the meat image as a basis for the graded meat, and stores code that causes a terminal in communication with the processor to provide an interface including a grade, an evaluation score, a meat image, a QR code, and visualization data.

10. In paragraph 8, A meat grading device, wherein the above grading information is set based on criteria for judging multiple pre-set national beef or pork gradings.

11. In paragraph 9, The grade of meat included in the above meat image includes at least one grade among multiple national grades, The above memory, when executed through the processor, causes the processor to: A meat grading device that stores a code that causes meat included in the meat image to be judged and provided based on criteria corresponding to the grade of the specific country when a request for grading of a specific country is received from a terminal connected to the processor.

12. In paragraph 9, A meat grading device, wherein the above-mentioned area is a meat area containing a certain portion or more of the protein contained in the meat image.

Images