Pet health assessment system based on AI analysis

Abstract

The invention provides a pet health assessment system based on AI analysis, and the system comprises a first analysis detection module which obtains pet representation information, and forms a preliminary assessment result according to the obtained information; the first analysis and detection module is a multi-modal model, and the pet representation information is obtained through interaction with a user; the user judges whether to start a second analysis detection module or not according to the preliminary evaluation result, obtains physiological information and sends the obtained physiological information to a first detection module; the management model is used for acquiring information of the first analysis and detection module and the second analysis and detection module, the management model is one given by a machine learning model, an AI comprehensive model or a stacking comprehensive model, then health management information is given, and management suggestions in pet care are given according to the health management information. The system can actively interact with a pet owner, obtain key representation information, analyze medical detection data through an AI model, and provide comprehensive health management suggestions.

Description

Technical Field

[0001] The present invention relates to the technical field of pet health management, and in particular to a pet health assessment system based on AI analysis. Background Art

[0002] In the existing technology, pet health diagnosis and disease prevention mainly rely on traditional testing methods such as veterinary expertise and experience. This process involves the analysis of multiple independent health test data, such as physiological indicators such as body temperature, heart rate, blood pressure, and the monitoring of pets' daily activities.

[0003] The above method is effective to a certain extent, but there is a problem of high dependence on professional judgment. The judgment and experience of veterinarians play a decisive role in pet health management, but the strong subjectivity brought about by this high degree of personal dependence leads to excessive subjectivity and inconsistency in health management, and thus low overall accuracy.

[0004] With the development of artificial intelligence, especially the application of multimodal large models and deep learning technology, pet health assessment can be more accurate and automated. However, existing technologies often lack sufficient interactivity and initiative when processing pet health data, and cannot effectively obtain key information from pet owners. At the same time, they also fail to make full use of advanced AI models to analyze complex medical test data. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a pet management system based on AI analysis, using multimodal large models and deep learning technology to achieve accurate assessment of pet health status. The system can actively interact with pet owners, obtain key pet appearance information, and analyze medical test data through advanced AI models to provide comprehensive health management suggestions.

[0006] In order to solve the above technical problems, the technical solutions provided by the present invention are as follows: Pet health assessment system based on AI analysis, including: A first analysis and detection module is provided for obtaining pet appearance information, wherein the first analysis and detection model forms a preliminary evaluation result based on the obtained information; The first analysis and detection module is a multimodal model, and the pet appearance information is obtained by interacting with the user; The user determines whether to start the second analysis and detection module based on the preliminary evaluation result, and the second analysis and detection module obtains physiological information and sends the obtained physiological information to the first detection module; A management model for obtaining information from the first analysis and detection module and the second analysis and detection module is provided, wherein the management model is one of a machine learning model, an AI comprehensive model or a stacked comprehensive model, and then health management information is provided, and management suggestions for pet care are provided based on the health management information.

[0007] Furthermore, it also includes the acquisition of pet appearance information, specifically, obtaining a first acquisition and analysis module containing M appearance parameter features and subjective medical conclusions through human-computer interaction, where M is a natural number greater than or equal to 1.

[0008] Furthermore, in the human-computer interaction mode, when the appearance parameter characteristics cannot be satisfied, the user is reminded to supplement the key parameters, and when the user cannot provide the key parameters, the user is prompted to start the second analysis and detection module according to the key parameters.

[0009] Furthermore, the first acquisition and analysis module includes a consultation and interaction module, and the consultation and interaction module is based on physiological information actively input and / or passively responded by the pet owner, including but not limited to basic information of the pet.

[0010] Furthermore, the basic information of the pet includes but is not limited to name, gender, age, vaccination information and historical medical information.

[0011] Furthermore, it also includes a file management module, which stores at least the basic information, beauty information, and deworming information of the pet, and updates it in time.

[0012] Furthermore, the multimodal model is used to obtain user input data in human-computer interaction, including text, images, existing medical records, existing medical test results or audio.

[0013] Furthermore, it also includes the acquisition of pet physiological information, specifically, a second acquisition and analysis module containing N medical parameter indicators and objective medical conclusions obtained through a plurality of detection units, wherein N is a natural number greater than or equal to 1.

[0014] Furthermore, the second acquisition and analysis module is provided with several AI models, and the several AI models respectively analyze the detection information obtained by the detection unit by calling VGG or DenseNet to determine whether it is abnormal.

[0015] Furthermore, the management model uses a pre-trained machine learning model, an AI comprehensive model or a stacked comprehensive model for analysis, optimizes the analysis results through weighting and feature selection, and then generates a management report containing health assessments and maintenance recommendations.

[0016] The present invention has the following beneficial effects: First, the present invention selects two analysis models and combines them with the final management model, combining subjective and objective data and analysis basis, so that the final conclusion is more objective and the health management suggestions given are more accurate, which is convenient for pet owners to accurately manage their pets.

[0017] Secondly, in the present invention, a multimodal large model and deep learning technology are used to achieve an accurate assessment of the health status of pets. The system can actively interact with pet owners to obtain key pet appearance information, and can also integrate the archive management system to obtain existing pet-related information. In addition, it can analyze existing test reports in a multimodal manner, and analyze medical test data through advanced AI models to provide comprehensive health management recommendations.

[0018] Again, the management system in the present invention first obtains the pet's appearance information, and then conducts in-depth analysis and detection based on the appearance, thus having a certain progressive relationship, avoiding high consumption problems caused by blind detection, and due to the existence of multiple analysis models, the accuracy, preventiveness and user participation of pet health management are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flowchart of the use of the pet health assessment system based on AI analysis provided by the present invention. DETAILED DESCRIPTION

[0020] The present invention is described in detail below in conjunction with the various embodiments shown in the accompanying drawings, but it should be noted that these embodiments are not limitations of the present invention, and any equivalent transformations or substitutions in functions, methods, or structures made by ordinary technicians in the field based on these embodiments are all within the scope of protection of the present invention.

[0021] The pet health assessment system based on AI analysis in the present invention mainly includes a first analysis and detection module, a second analysis and detection module and a management module. The three modules are introduced separately below.

[0022] First, the first analysis and detection module.

[0023] The purpose of this module is to obtain the pet's appearance information and form a preliminary evaluation result based on the obtained information; at this time, the acquisition of the pet's appearance information is specifically to obtain a first acquisition and analysis module containing M appearance parameter features and subjective medical conclusions through human-computer interaction, where M is a natural number greater than or equal to 1. At this time, the appearance parameters include the pet dog's hair, skin condition, and mental state, etc.

[0024] In this embodiment, the first analysis and detection module is a multimodal model, and the pet appearance information is obtained through interaction with the user; the first acquisition and analysis module includes a consultation interaction module, and the consultation interaction module is based on physiological information actively input and / or passively responded by the pet owner, including but not limited to the basic information of the pet.

[0025] The pet's basic information includes, but is not limited to, name, gender, age, vaccination information and historical medical information, and a preliminary rough judgment can be made by comparing the basic information with some corresponding physiological information.

[0026] In order to make full use of the existing technology, a file management module is also included, which stores at least the basic information, beauty information, and deworming information of the pet, and is updated in time. The file management module provided in this embodiment includes files uploaded in human-computer interaction, or information obtained after association with other management systems.

[0027] Specifically, the multimodal model is used to obtain user input data in human-computer interaction, including text, images, existing medical records, existing medical test results or audio.

[0028] In actual use, the first analysis and detection module: The core of this module is a large multimodal model that has the ability to interact with pet owners and obtain the pet's appearance information, such as behavior, appearance, etc. The large multimodal model analyzes this information to determine whether it is sufficient for an effective health assessment. If the model determines that the information is insufficient, it will actively identify and ask the pet owner about key but missing parameters, such as the pet's eating habits, daily activity, etc. These parameters are crucial for accurately assessing the health of the pet. If these key parameters cannot be obtained through interaction with the pet owner, in order to ensure the accuracy and reliability of the evaluation results, the system will declare that no further judgment can be made and terminate the evaluation attempt.

[0029] In actual use, the first analysis and detection module at this time is supported by a multimodal large model that has the ability to communicate with people. The large model itself can determine whether the parameters are sufficient. In the case of insufficient parameters, the large model will actively ask the pet owner one by one for the parameters that are likely to be obtained from the pet owner. If the key parameters that can only be obtained from the pet owner cannot be obtained in this way, in order to ensure the quality of the conclusion, it can only be announced that no further judgment can be made and the attempt will be terminated. For other parameters that need to be obtained through medical testing, the first analysis and detection module will delegate the task to the second analysis and detection module.

[0030] The specific working process of the first analysis and detection module is as follows: Obtain pet appearance information, determine the adequacy of the information, and actively inquire about key parameters when the information is insufficient. If it cannot be obtained, terminate the analysis and evaluation and then transfer it to the second analysis and detection module.

[0031] In the human-computer interaction mode, when the appearance parameter characteristics cannot be met, the user is reminded to supplement the key parameters, and when the user cannot provide the key parameters, the user is prompted to start the second analysis and detection module according to the key parameters. Specifically, the user determines whether to start the second analysis and detection module based on the preliminary evaluation results, and the second analysis and detection module obtains physiological information and sends the obtained physiological information to the first detection module.

[0032] Secondly, the second analysis and detection module.

[0033] In this embodiment, the second analysis and detection module is mainly used to obtain physiological information and send the obtained physiological information to the first detection module. Specifically, it is a second acquisition and analysis module that obtains N medical parameter indicators and has objective medical conclusions through a number of detection units, where N is a natural number greater than or equal to 1.

[0034] The second acquisition and analysis module is provided with several AI models, and the several AI models respectively analyze the detection information obtained by the detection unit by calling VGG or DenseNet to determine whether it is abnormal.

[0035] This module focuses on obtaining supplementary physiological information of pets, which usually needs to be obtained through professional medical testing methods. Specifically, the second analysis and detection module uses advanced AI analysis models, which can be set to multiple types, such as VGG and DenseNet, to conduct in-depth analysis of the pet's lung X-rays or head CT scans. These models can identify abnormal patterns in images by learning a large amount of medical imaging data, thereby helping to discover potential health problems. In the specific process, the second analysis and detection module first receives a request from the first analysis and detection module to clarify the type of physiological parameters that need to be obtained. Subsequently, the module coordinates the corresponding medical testing equipment to perform the necessary image acquisition. After the acquisition is completed, the image data will be sent to the pre-trained VGG or DenseNet model for analysis. These models will output analysis results, including possible abnormal areas and corresponding probability scores. These results will then be returned to the first analysis and detection module for comprehensive health assessment based on all information.

[0036] Again, management model.

[0037] A management model for obtaining information from the first analysis and detection module and the second analysis and detection module is provided, wherein the management model is one of a machine learning model, an AI comprehensive model or a stacked comprehensive model, and then health management information is provided, and management suggestions for pet care are provided based on the health management information.

[0038] In this embodiment, the management model uses a pre-trained machine learning model, an AI comprehensive model, or a stacked comprehensive model for analysis, optimizes the analysis results through weighting and feature selection, and then generates a management report containing health assessments and maintenance recommendations. The management model integrates the information of the first and second analysis and detection modules, and uses a machine learning model, an AI comprehensive model, or a stacked comprehensive model to provide health management information and pet maintenance recommendations.

[0039] In this embodiment, the management model receives inputs from two analysis modules, including the pet's appearance information, physiological parameters, and medical imaging results obtained through AI model analysis. These input data will be fed into one or more pre-trained machine learning models, which may include decision trees, linear regression neural networks (such as multi-layer perceptrons MLP), or multimodal large models. During the implementation process, the management model first pre-processes the received data to ensure the consistency of the data format and content. Subsequently, the model selects the appropriate machine learning algorithm for analysis based on the type and characteristics of the input data. For example, for continuous physiological parameters, a linear regression model may be used; for classification problems, such as disease diagnosis, a decision tree or support vector machine may be used. During the analysis process, the model considers the relevance and weight of all input data to ensure the accuracy and comprehensiveness of the health assessment.

[0040] In this embodiment, the classification and processing of information are also included. In the management model, the information is divided into two categories: information obtained by the first detection module and information that must be obtained with the help of specific tests to ensure the integrity and accuracy of the information. In order to ensure the integrity and accuracy of the pet health assessment, the system divides the information into two categories: one is information obtained by the first detection module, and the other is information that must be obtained with the help of specific tests. The first detection module, as a multimodal large model, can obtain the pet's appearance information, such as behavior, appetite, activity, etc., through interaction with the pet owner. When obtaining this information, the large model will determine whether the parameters are sufficient, and if the parameters are insufficient, it will actively ask the pet owner for key parameters that may be missing. If these key parameters cannot be obtained through interaction with the pet owner, the system will declare that no further judgment can be made and terminate the attempt. For those physiological parameters that can only be obtained through medical testing, the first detection module will delegate the task to the second detection module. The second detection module is responsible for obtaining supplementary physiological information, such as blood tests, urine analysis, X-ray analysis, etc., and sending this information back to the first detection module. When processing medical images, the second detection module may call specific AI models, such as VGG, DenseNet, etc., to conduct in-depth analysis of the images to detect potential abnormalities. During the implementation process, the first detection module will first try to obtain all possible information through interaction with the pet owner. If certain key parameters cannot be obtained in this way, the system will clearly inform the pet owner which specific medical tests are needed. Once the second detection module provides this supplementary information, the first detection module will integrate it into the existing information to ensure the integrity of the information. Subsequently, the management model will receive this integrated information and use the machine learning model for in-depth analysis to ultimately generate health management information and maintenance recommendations for the pet.

[0041] Finally, health assessment recommendations.

[0042] In this embodiment, based on all the collected information, the system will conduct a comprehensive analysis through the AI ​​model to draw conclusions about the health status of the pet and provide targeted health management suggestions. This process involves two main steps: health status assessment and health management suggestion generation.

[0043] In the first step, health status assessment is achieved by inputting all the information collected by the first detection module and the second detection module into a comprehensive AI model. This comprehensive AI model can be a machine learning model, an AI comprehensive model, or a stacked comprehensive model. The model analyzes a variety of input data, including pet appearance information (such as behavior, appetite, activity level, etc.) and physiological parameters (such as blood test results, urine analysis, X-ray analysis, etc.). For example, the model may use a decision tree or a linear regression neural network trained by supervised learning (such as a multilayer perceptron MLP) to process this data to identify potential health problems.

[0044] In the second step, health management recommendations are generated based on the results of the health status assessment. The system will provide specific maintenance recommendations based on the identified health problems. For example, if the assessment results show that the pet has kidney problems, the system will recommend adjusting the diet, reducing protein intake, and performing regular kidney function tests. If the pet has bone problems, the system may recommend increasing calcium intake and exercising appropriately.

[0045] This embodiment also discloses a pet health assessment method based on AI analysis, which specifically includes the following steps: First, the pet owner enters the pet's basic information and appearance information through the system interface. After receiving this information, the multimodal large model performs a preliminary analysis to determine whether more parameters are needed. If necessary, the model will prompt the pet owner to provide the missing key parameters through the system interface. For example, if the pet has recently lost its appetite, the model may ask about specific dietary changes. If the pet owner is unable to provide this information, the system will display a prompt that "further evaluation is not possible" and suggest that the pet owner seek help from a professional veterinarian. For parameters that require professional testing, the system will arrange for the second analysis and detection module to perform the corresponding test and feed the results back to the first analysis and detection module so that all information can be combined to obtain the final health assessment result.

[0046] Secondly, if a lung X-ray or head CT scan is indicated, the medical device will generate the corresponding image data after the pet receives the lung X-ray or head CT scan. These data are transmitted to the second analysis and detection module through the network. After receiving the data, the module automatically calls the pre-configured VGG or DenseNet model for analysis. The model will carefully inspect each area in the image and mark any suspicious abnormalities. After the analysis is completed, the module will generate a detailed report, including a description of the abnormal area and a probability score. This report will then be sent back to the first analysis and detection module for comprehensive evaluation in combination with other information. For example, if the model finds a high-probability shadow area in the lung X-ray, the report will recommend further diagnostic or treatment measures. In this way, the second analysis and detection module provides key physiological information support for the health assessment of the pet.

[0047] Again, after receiving the data from the first and second analysis and detection modules, the management model first cleans and formats the data, removes irrelevant or erroneous data, and converts all data into a format acceptable to the model. Then, the model selects the most appropriate machine learning algorithm for analysis based on the characteristics of the data and historical training experience. During the analysis process, the model considers the importance of each input data and optimizes the analysis results through weighting and feature selection. Finally, the model generates a report containing health assessment and maintenance recommendations. For example, if the model analysis results show that the pet has a higher risk of obesity, the report will recommend adjusting the diet and increasing exercise. In this way, the management model provides pet owners with personalized maintenance guidance based on comprehensive data analysis. At this point, based on all the information collected, the system conducts a comprehensive analysis through the AI ​​model, draws conclusions about the health status of the pet, and provides targeted health management recommendations. This process involves two main steps: health status assessment and health management recommendation generation.

[0048] In the case of a pet health assessment consultation, the details are as follows: If a pet owner reports that his pet has recently lost its appetite, the first analysis and detection module will first ask a series of relevant questions, such as the pet's activity level and whether it has symptoms such as vomiting or diarrhea. If this information is not enough for an accurate assessment, the system will recommend blood tests and urine analysis. Once the second analysis and detection module provides these test results, the first analysis and detection module will integrate this information and send it to the management model. The management model will select the appropriate machine learning algorithm for analysis based on all the input data, and finally generate a report containing health assessments and maintenance recommendations. For example, if the analysis results show that the pet has kidney problems, the report will recommend adjusting the diet and performing regular kidney function tests. In this way, the system ensures the comprehensiveness and accuracy of the pet's health assessment and provides pet owners with a health management plan based on scientific data.

[0049] In this embodiment, regarding the management model, first, health status assessment is achieved by inputting all the information collected by the first analysis and detection module and the second analysis and detection module into a comprehensive AI model. This comprehensive AI model can be one of a machine learning model, an AI comprehensive model, or a stacked comprehensive model. The model analyzes a variety of input data, including pet appearance information (such as behavior, appetite, activity level, etc.) and physiological parameters (such as blood test results, urine analysis, X-ray analysis, etc.). For example, the model may use a decision tree or a linear regression neural network trained by supervised learning (such as a multilayer perceptron MLP) to process this data to identify potential health problems.

[0050] Secondly, health management recommendations are generated based on the results of health status assessment. The system will provide specific maintenance recommendations based on the identified health problems. For example, if the assessment results show that the pet has kidney problems, the system will recommend adjusting the diet, reducing protein intake, and performing regular kidney function tests. If the pet has bone problems, the system may recommend increasing calcium intake and exercising appropriately.

[0051] See attached Figure 1 As shown, using the pet health assessment system based on AI analysis in the present invention, the assessment method steps are as follows: The following is mainly based on the description of the interaction process between a pet owner and an AI health manager in the pet health assessment system based on AI analysis in this embodiment.

[0052] The historical data, or pet file, is as follows: 6.5-year-old British dog, male (castrated), weight 35Kg, swelling under the nail of the fourth toe on the left hind leg, pain relief, nail is firm, and the surface of the toe is ulcerated and scabby.

[0053] Pet owner: My dog ​​is male (castrated), weighing 35Kg, dewormed, fully vaccinated, with good appetite, normal spirit, history of skin disease, sparse hair and pigmentation on the left side of the abdomen about 10cmx10cm, severe obesity, visible mucous membrane pink and mild tartar, red ears and four paw pads. The fourth toe of the left hind leg is swollen and painful, with missing nails, and the surface of the toe is ulcerated with scabs. The lesion was found 3 days ago, and the examination at other hospitals was diagnosed as toe trauma. After 2 days of treatment, it has not been relieved.

[0054] Health Manager: Based on the information you provided, regarding the swelling, pain, nail loss, and surface ulcers with scabs under the nail of the fourth toe of the left hind leg of the English bulldog, possible diagnoses include: 1. Infection: A bacterial or fungal infection may cause redness, pain, and ulcer formation on the toe. 2. Trauma: Direct impact, cutting, or squeezing may cause trauma, which may lead to infection. 3. Foreign body: The toe may step on or insert a foreign body when walking, causing inflammation and infection. 4. Tumor: Although the possibility is low, it cannot be ruled out that the swelling is caused by malignant or benign tumors. Treatment is recommended in the following order: Continue to keep the wound clean and observe for signs of suppuration or increased discharge. If necessary, change antibiotics or antifungal drugs, based on the previous treatment effect and the advice of the veterinarian. If the symptoms do not improve or worsen, further diagnostic tests should be considered, such as routine blood tests, X-rays to rule out fractures or foreign bodies, and possible biopsy (biopsy) to rule out tumors. For severe obesity problems, a weight loss plan and diet adjustment are required, because excessive weight will increase the burden on the joints and affect healing. Please continue to monitor the status of the toe based on the information you provided.

[0055] At this point, the function of the first analysis and detection module is completed, and subsequent reports can be completed by the second analysis and detection module, or supplemented and improved by existing known information.

[0056] In the above, the role of the AI ​​health manager is to consider several possible causes of the disease based on text and pictures, provide treatment strategies based on currently known conditions, and recommend further tests for confirmation, including blood routine, X-rays, and biopsies.

[0057] User: Input Figure 1 Blood test report in; AI Health Manager: If the blood test results are normal, some systemic diseases or metabolic problems can be ruled out. The next suggestion is: consider imaging tests, such as X-rays or ultrasounds, to evaluate the bone structure and check for foreign matter.

[0058] In the above, the role of the AI ​​health manager is to basically rule out the possibility of systemic diseases based on the results of the blood routine test, and determine in this case that a concentrated examination of the affected area is required, starting with the X-ray or B-ultrasound results.

[0059] User: Input X-ray photos (here it is achieved by sending the images after X-ray examination); AI Health Manager: X-ray examination showed soft tissue swelling of the fourth toe of the left hind limb and a mass of uniform density at the distal end, which may be one of the following: 1. Soft tissue tumor: benign or malignant, a biopsy is required to determine the type of tumor. 2. Chronic infection: Long-term infection response may cause local tissue hyperplasia. 3. Granuloma: Inflammatory tissue related to infection or foreign body reaction. The treatment recommendations are as follows: Perform a local biopsy to confirm whether it is a tumor and determine whether it is benign or malignant.

[0060] In the above, the role of the AI ​​health manager is to exclude the possibility of trauma, foreign matter, etc. based on the X-ray film, and then focus on a limited number of possibilities, and recommend a local biopsy to determine the cause. User: Input photos of cell staining report, as well as skin tissue examination report and photos.

[0061] Health Manager: Assess the existence of health problems, specifically the results of tissue pathology examination may show poorly differentiated squamous cell carcinoma, which is a malignant tumor. The specific treatment plan should be determined by a professional veterinarian based on clinical experience and the specific location, size and whether the tumor has metastasized. Common treatment options may include: 1. Surgery If the tumor is confined to one area and has not metastasized, surgical resection may be the first choice. 2. Chemotherapy or radiotherapy: In cases where complete surgical resection is not possible or metastasis has already occurred, chemotherapy or radiotherapy may be considered to control tumor growth and relieve symptoms. 3. Supportive treatment: including analgesics, anti-inflammatory drugs and measures to improve the quality of life. Please discuss the treatment plan with your veterinarian as soon as possible, and consider consulting an oncologist. Early intervention and comprehensive treatment are very important for the management of your pet's condition. At the same time, maintain regular checkups to monitor changes in the tumor and the health of other parts of the body.

[0062] In the above, the role of the AI ​​health manager is to obtain the biopsy results. The multimodal large model automatically calls the corresponding professional medical models (blood routine analysis model, X-ray analysis model, cell staining image analysis model, etc.) to conduct detailed analysis and interpretation of the above related test results and obtain the analysis results (the blood routine model can point out which indicators are abnormal and the corresponding symptoms; the X-ray analysis model and the cell staining image analysis model can clearly point out the abnormal parts and causes in the medical images). Afterwards, based on the information provided by the user and the analysis results of the professional medical model, the multimodal large model basically determines that the pet suffers from poorly differentiated squamous cell carcinoma and provides feasible treatment options.

[0063] After the consultation, the AI ​​health manager gave the following comprehensive conclusion and evaluation report: Based on the above examination results and analysis, it is preliminarily diagnosed as poorly differentiated squamous cell carcinoma. This is a malignant tumor and treatment measures need to be taken as soon as possible. Please work closely with your veterinarian to discuss the treatment options suitable for your pet, including surgery, chemotherapy or radiotherapy. At the same time, pay attention to monitoring your pet's overall health and quality of life, and conduct regular follow-up examinations.

[0064] Here, the AI ​​health manager gives a concise summary of the consultation, and then simply gives suggestions, as well as some possible treatment options that can be found on Baidu.

[0065] In this embodiment, before going to a professional medical institution, an assessment is first performed through the pet health assessment system. If any problems are found, the owner is promptly reminded to take the pet to the hospital for treatment, thereby reducing the risk of subsequent deterioration.

[0066] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0067] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A pet health assessment system based on AI analysis, characterized by: include, A first analysis and detection module is provided for obtaining pet appearance information, wherein the first analysis and detection model forms a preliminary evaluation result based on the obtained information; The first analysis and detection module is a multimodal model, and the pet appearance information is obtained by interacting with the user; The user determines whether to start the second analysis and detection module based on the preliminary evaluation result, and the second analysis and detection module obtains physiological information and sends the obtained physiological information to the first detection module; A management model for obtaining information from the first analysis and detection module and the second analysis and detection module is provided, wherein the management model is one of a machine learning model, an AI comprehensive model or a stacked comprehensive model, and then health management information is provided, and management suggestions for pet care are provided based on the health management information.

2. The pet management system based on AI analysis according to claim 1, characterized in that: It also includes the acquisition of pet appearance information, specifically, obtaining a first acquisition and analysis module containing M appearance parameter features and subjective medical conclusions through human-computer interaction, where M is a natural number greater than or equal to 1.

3. The pet management system based on AI analysis according to claim 2, characterized in that: In the human-computer interaction mode, when the appearance parameter characteristics cannot be met, the user is reminded to supplement the key parameters, and when the user cannot provide the key parameters, the user is prompted to start the second analysis and detection module according to the key parameters.

4. The pet management system based on AI analysis according to claim 2, characterized in that: The first acquisition and analysis module includes a consultation and interaction module, and the consultation and interaction module is based on the physiological information actively input and / or passively responded by the pet owner, including but not limited to the basic information of the pet.

5. The pet management system based on AI analysis according to claim 4 is characterized in that: The pet's basic information includes but is not limited to name, gender, age, vaccination information and historical medical information.

6. The pet management system based on AI analysis according to claim 5, characterized in that: It also includes a file management module, which stores at least the pet's basic information, beauty information, and deworming information, and updates them in a timely manner.

7. The pet management system based on AI analysis according to claim 6, characterized in that: The multimodal model is used to obtain user input data in human-computer interaction, including text, images, existing medical records, existing medical test results or audio.

8. The pet management system based on AI analysis according to claim 1, characterized in that: It also includes the acquisition of pet physiological information, specifically a second acquisition and analysis module containing N medical parameter indicators and objective medical conclusions obtained through a number of detection units, where N is a natural number greater than or equal to 1.

9. The pet management system based on AI analysis according to claim 7, characterized in that: The second acquisition and analysis module is provided with several AI models, and the several AI models respectively analyze the detection information obtained by the detection unit by calling VGG or DenseNet to determine whether it is abnormal.

10. The pet management system based on AI analysis according to claim 1, characterized in that: The management model uses a pre-trained machine learning model, an AI comprehensive model, or a stacked comprehensive model for analysis, optimizes the analysis results through weighting and feature selection, and then generates a management report containing health assessments and maintenance recommendations.

Images