KR-20260067087-A - Apparatus and method for diagnosing fish diseases using artificial intelligence

KR20260067087AKR 20260067087 AKR20260067087 AKR 20260067087AKR-20260067087-A

Abstract

An apparatus and method for diagnosing fish diseases using artificial intelligence are disclosed. The method for diagnosing fish diseases using artificial intelligence comprises the steps of: generating a fish disease symptom detection model through learning using fish disease diagnosis data in which fish images and disease information are mapped according to fish diseases; generating a fish disease classification model through learning using fish disease symptom data generated from the fish disease diagnosis data; receiving a fish image of a fish to be tested; inputting the input fish image into the fish disease symptom detection model to extract symptom information of the fish to be tested; and inputting the extracted symptom information into the fish disease classification model to derive a disease classification result of the fish to be tested.

Inventors

최한석
정종호
김찬진

Assignees

빛가람정보주식회사

Dates

Publication Date: 20260512
Application Date: 20241105

Claims (10)

In a method for diagnosing fish diseases using artificial intelligence performed by a fish disease diagnostic device, A step of generating a fish disease symptom detection model through training using fish disease diagnostic data in which fish images and disease information are mapped by fish disease; A step of generating a fish disease classification model through learning using fish disease symptom data generated from the above fish disease diagnosis data; A step of receiving an image of the fish to be inspected; A step of inputting the above-mentioned fish image into the above-mentioned fish disease symptom detection model to extract symptom information of the above-mentioned fish to be examined; and A method for diagnosing fish diseases using artificial intelligence, comprising the step of inputting the extracted symptom information into the fish disease classification model to derive a disease classification result of the fish to be examined.
In paragraph 1, The step of generating the above-mentioned fish disease symptom detection model is, A step of receiving the above fish disease diagnosis data; A step of generating fish disease symptom data from the above-mentioned input fish disease diagnosis data, wherein the fish image and labeling data recording the fish symptoms form a pair; A step of preprocessing the fish disease symptom data generated above; A step of generating training data by classifying the above-mentioned preprocessed fish disease symptom data through a training data distributor (Train/Valid/Test Dispenser); and A method for diagnosing fish diseases using artificial intelligence, characterized by including the step of generating a fish disease symptom detection model by training the above-mentioned training data with YOLO (You Only Look Once).
In paragraph 2, The step of generating the above fish disease symptom data is, A method for diagnosing fish diseases using artificial intelligence, characterized by generating fish disease symptom data by tokenizing and extracting fish image and labeling data from the above-mentioned fish disease diagnosis data.
In paragraph 2, The above preprocessing step is, A method for diagnosing fish diseases using artificial intelligence, characterized by performing operations of data extraction, addition, removal, and division on the fish disease symptom data according to user input, and performing a verification operation using a preset algorithm.
In paragraph 2, The step of generating the above fish disease classification model is, A step of organizing and structuring the above fish disease symptom data, configured by fish image, by individual; A step of generating fish disease text data by converting the data generated through the above structuring into a text form; and A method for diagnosing fish diseases using artificial intelligence, characterized by including the step of generating a fish disease classification model by training the above-mentioned fish disease text data with an SGD Classifier and a RandomForest Classifier.
In paragraph 5, The above structuring step is, A method for diagnosing fish diseases using artificial intelligence, characterized by extracting images representing object information and categories including disease and symptom information from a COCO (Common Objects in Context) dataset for the labeling data of the fish disease symptom data, and using the extracted images and categories to generate a JSON object in which the object information is the key and a multi-digit code representing the disease and symptoms of the object is the value.
In paragraph 6, In order to convert the above JSON object into learnable data, a Data Frame is created in which, among the above multi-digit codes, the codes representing symptoms are used as columns, the codes representing diseases are used as label values, and object information is used as key values. A method for diagnosing fish diseases using artificial intelligence, characterized in that the above-mentioned heat information value is set to indicate the presence or absence of symptoms for each individual (Index).
In Paragraph 7, The step of generating the above fish disease text data is, A method for diagnosing fish diseases using artificial intelligence, characterized by generating fish disease text data by performing a label encoder process that converts data into a numerical form.
In paragraph 8, The above fish disease text data is structured to have three types of data: an entity name, disease information, and a symptom status table, and A method for diagnosing fish diseases using artificial intelligence, characterized in that the above-mentioned individual name represents a unique index value of a fish, the above-mentioned disease information refers to disease information included in the above-mentioned fish disease diagnosis data, and the above-mentioned symptom status table refers to a table containing data converted into numbers during the above-mentioned label conversion process.
In a fish disease diagnosis device using artificial intelligence, Memory for storing instructions; and A processor that executes the above instructions, The above command is, A step of generating a fish disease symptom detection model through training using fish disease diagnostic data in which fish images and disease information are mapped by fish disease; A step of generating a fish disease classification model through learning using fish disease symptom data generated from the above fish disease diagnosis data; A step of receiving an image of the fish to be inspected; A step of inputting the above-mentioned fish image into the above-mentioned fish disease symptom detection model to extract symptom information of the above-mentioned fish to be examined; and A fish disease diagnosis device using artificial intelligence, characterized by performing a fish disease diagnosis method comprising the step of inputting the extracted symptom information into the fish disease classification model to derive a disease classification result of the fish to be examined.

Description

Apparatus and method for diagnosing fish diseases using artificial intelligence The present invention relates to a device and method for diagnosing fish diseases using artificial intelligence. Aquaculture accounts for a significant portion of domestic fisheries production, and flounder, in particular, is a major fish species with high production volume and economic value. However, flounder is susceptible to various diseases such as vibriosis and Edwardsiella, and failure to take appropriate measures when diseases occur can result in massive economic losses. Conventional methods for diagnosing flounder diseases have primarily relied on visual observation by skilled experts. However, this approach is susceptible to the subjective judgment of experts, and the diagnosis rate varies significantly depending on their skill and condition, making early diagnosis and treatment difficult. In contrast, artificial intelligence technologies, including deep learning and machine learning, can provide consistent and objective diagnostic results regardless of human condition. Furthermore, because they can rapidly analyze large volumes of data and detect even minute patterns to identify early symptoms that humans might otherwise miss, they are suitable for the early diagnosis and treatment of diseases. Deep learning technology demonstrates outstanding performance in the field of image recognition and analysis and is being utilized in various fields such as medicine and agriculture. While research on classifying types of diseases and early diagnosis through deep learning-based image analysis is actively underway, research on applying deep learning to the diagnosis of flounder diseases is still lacking. FIG. 1 is a flowchart schematically illustrating a method for diagnosing fish diseases using artificial intelligence performed by a fish disease diagnostic device according to an embodiment of the present invention. FIG. 2 is a flowchart schematically illustrating the detailed steps of step S110 of FIG. 1. FIG. 3 is a flowchart schematically illustrating the detailed steps of step S120 of FIG. 1. FIGS. 4 to 10 are drawings for explaining a method for diagnosing fish diseases using artificial intelligence according to an embodiment of the present invention of FIG. 1. FIG. 11 is a diagram schematically illustrating the configuration of a fish disease diagnosis device using artificial intelligence according to an embodiment of the present invention. As used in this specification, singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "composed" or "comprising" should not be interpreted as necessarily including all of the various components or steps described in the specification, and should be interpreted as meaning that some of the components or steps may be excluded, or that additional components or steps may be included. Furthermore, terms such as "...part," "module," etc., as used in the specification refer to a unit that processes at least one function or operation, which may be implemented in hardware or software, or a combination of hardware and software. Hereinafter, various embodiments of the present invention will be described in detail with reference to the attached drawings. FIG. 1 is a flowchart schematically illustrating a method for diagnosing fish diseases using artificial intelligence performed by a fish disease diagnostic device according to an embodiment of the present invention, FIG. 2 is a flowchart schematically illustrating a detailed step of step S110 of FIG. 1, FIG. 3 is a flowchart schematically illustrating a detailed step of step S120 of FIG. 1, and FIG. 4 to 10 are drawings for explaining the method for diagnosing fish diseases using artificial intelligence according to an embodiment of the present invention of FIG. 1. Hereinafter, the method for diagnosing fish diseases using artificial intelligence according to an embodiment of the present invention will be described with reference to FIG. 2 to 10. In step S110, the fish disease diagnostic device generates a fish disease symptom detection model through learning using fish disease diagnostic data in which fish images and disease information are mapped by fish disease. Hereinafter, with reference to FIG. 2, the detailed steps of step S110 will be explained. In step S111, the fish disease diagnostic device receives fish disease diagnostic data in which fish images and disease information are mapped according to fish diseases. Here, the disease information may include the disease name, diseased part, disease symptoms, etc. For example, fish disease diagnostic data can be obtained as shown in Fig. 4. Fig. 4 shows flounder disease diagnostic data for bacterial diseases that frequently occur in aquaculture farms, such as streptococcal disease, vibriosis, gliding bacterial disease, and Edwardsiella, viral diseases such as viral hemorrhagic septicemia (VHSV), and parasitic diseases s