Pet disease intelligent diagnosis and treatment data processing system and method

Abstract

The invention provides a pet disease intelligent diagnosis and treatment data processing system and method, relates to the field of pet diagnosis and treatment, and adopts a semantic analysis technology based on deep learning to perform semantic coding on pet symptom expression text description to obtain pet symptom expression semantic features. Meanwhile, semantic coding is carried out on pet age, variety and past medical history information to obtain pet diagnosis auxiliary information semantic features, and pet symptom expression semantic features and pet diagnosis auxiliary information semantic features are comprehensively considered, so that pet disease diagnosis is realized. Therefore, the accuracy of pet disease diagnosis can be improved, and timely pet disease diagnosis and treatment help can be provided in non-working time.

Description

Technical Field

[0001] The present application relates to the field of pet diagnosis and treatment, and more specifically, to a pet disease intelligent diagnosis and treatment data processing system and method. Background Art

[0002] In today's society, pets have become an indispensable and important member of many families. They provide people with companionship and warmth, and people also pay special attention to the health of their pets. However, the traditional online consultation model has many problems in practical application.

[0003] To a large extent, traditional online consultations rely mainly on the personal experience of veterinarians and the limited cases they have accumulated in their careers. Veterinarians rely on their own professional knowledge and past experience to judge the pet's condition, but since everyone's experience has certain limitations and the experience of different veterinarians is also different, this may inevitably lead to a certain degree of subjectivity in the diagnosis. At the same time, the number of cases accumulated by veterinarians is limited after all, and it is difficult to cover all possible conditions, which further limits the accuracy of the diagnosis.

[0004] In addition, traditional online consultations are also limited by time and veterinarians’ energy. During daily working hours, veterinarians need to deal with multiple consultations at the same time and may not be able to respond to every question in a timely manner. In the middle of the night and during holidays, veterinarians also need to rest, and pet owners’ inquiries are often not responded to in a timely manner. This means that pet owners cannot get timely professional help when their pets suddenly become ill, and can only wait anxiously or seek other ways that may not be professional, which undoubtedly brings potential risks to the health of their pets.

[0005] Therefore, an optimized pet disease intelligent diagnosis and treatment data processing solution is needed. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a pet disease intelligent diagnosis and treatment data processing system and method, which adopts semantic analysis technology based on deep learning, obtains the semantic features of pet symptom manifestations by semantically encoding the text description of pet symptom manifestations, and obtains the semantic features of pet diagnosis auxiliary information by semantically encoding the pet's age, breed, and previous medical history information, and comprehensively considers the semantic features of pet symptom manifestations and the semantic features of pet diagnosis auxiliary information, so as to realize the diagnosis of pet diseases. In this way, the accuracy of pet disease diagnosis can be improved, and timely pet disease diagnosis and treatment assistance can be provided during non-working hours.

[0007] According to one aspect of the present application, a pet disease intelligent diagnosis and treatment data processing system is provided, which includes:

[0008] The pet diagnosis and treatment related data collection module is used to obtain text descriptions of pet symptoms and information about the pet's age, breed, and medical history;

[0009] A pet symptom data semantic encoding module, used for semantically encoding the pet symptom text description to obtain a pet symptom semantic association feature vector;

[0010] A pet disease auxiliary information semantic coding module is used to semantically encode the pet's age, breed, and past medical history information to obtain a pet disease diagnosis auxiliary semantic coding feature vector;

[0011] A semantic coding feature fusion module, used for fusing the pet symptom manifestation semantic association feature vector and the pet disease diagnosis auxiliary semantic coding feature vector to obtain a pet disease diagnosis semantic feature vector;

[0012] The pet diagnosis and treatment result generating module is used to obtain the diagnosis result of the pet's disease based on the pet disease diagnosis semantic feature vector.

[0013] According to another aspect of the present application, a method for processing pet disease intelligent diagnosis and treatment data is provided, which includes:

[0014] Obtain text descriptions of pet symptoms and information about the pet's age, breed, and medical history;

[0015] Performing semantic encoding on the pet symptom manifestation text description to obtain a pet symptom manifestation semantic association feature vector;

[0016] Semantically encoding the pet's age, breed, and past medical history information to obtain a pet disease diagnosis-aided semantic encoding feature vector;

[0017] Fusion of the pet symptom manifestation semantic association feature vector and the pet disease diagnosis auxiliary semantic coding feature vector to obtain a pet disease diagnosis semantic feature vector;

[0018] Based on the pet disease diagnosis semantic feature vector, a diagnosis result of the pet disease is obtained.

[0019] Compared with the prior art, the pet disease intelligent diagnosis and treatment data processing system and method provided by the present application adopts a semantic analysis technology based on deep learning, obtains the semantic features of pet symptom manifestations by semantically encoding the text description of pet symptom manifestations, and obtains the semantic features of pet diagnosis auxiliary information by semantically encoding the pet age, breed, and past medical history information, and comprehensively considers the semantic features of pet symptom manifestations and the semantic features of pet diagnosis auxiliary information, thereby realizing the diagnosis of pet diseases. In this way, the accuracy of pet disease diagnosis can be improved, and timely pet disease diagnosis and treatment assistance can be provided during non-working hours. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0021] Figure 1 This is a system block diagram of a pet disease intelligent diagnosis and treatment data processing system according to an embodiment of the present application.

[0022] Figure 2 This is a block diagram of a semantic encoding module for pet symptom manifestation data in a pet disease intelligent diagnosis and treatment data processing system according to an embodiment of the present application.

[0023] Figure 3 4 is a block diagram of a semantic encoding module for pet disease auxiliary information in a pet disease intelligent diagnosis and treatment data processing system according to an embodiment of the present application.

[0024] Figure 4 This is a block diagram of a pet disease diagnosis semantic feature optimization unit in a pet disease intelligent diagnosis and treatment data processing system according to an embodiment of the present application.

[0025] Figure 5 This is a flow chart of a method for processing data for intelligent diagnosis and treatment of pet diseases according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0027] As mentioned in the above background technology, traditional online consultations rely heavily on the personal experience of veterinarians and the limited cases they have accumulated, which inevitably leads to certain subjectivity and limitations in diagnosis. In addition, traditional online consultations are limited by time and veterinarians' energy, and responses are often not timely, especially late at night and during holidays, which prevents pet owners from getting timely professional help. Therefore, an optimized intelligent diagnosis and treatment data processing solution for pet diseases is expected.

[0028] In recent years, deep learning and neural networks have been widely used in computer vision, natural language processing, text signal processing and other fields. In addition, deep learning and neural networks have also demonstrated a level close to or even beyond that of humans in image classification, object detection, semantic segmentation, text translation and other fields. The development of deep learning and neural networks provides new solutions and solutions for the intelligent diagnosis and treatment of pet diseases.

[0029] Figure 1 FIG. 1 is a system block diagram of a pet disease intelligent diagnosis and treatment data processing system according to an embodiment of the present application. Figure 1 As shown, in the pet disease intelligent diagnosis and treatment data processing system 100, it includes: a pet diagnosis and treatment related data collection module 110, which is used to obtain the text description of pet symptom manifestations and the pet's age, breed, and previous medical history information; a pet symptom manifestation data semantic encoding module 120, which is used to semantically encode the pet symptom manifestation text description to obtain the pet symptom manifestation semantic association feature vector; a pet disease auxiliary information semantic encoding module 130, which is used to semantically encode the pet's age, breed, and previous medical history information to obtain a pet disease diagnosis auxiliary semantic encoding feature vector; a semantic encoding feature fusion module 140, which is used to fuse the pet symptom manifestation semantic association feature vector and the pet disease diagnosis auxiliary semantic encoding feature vector to obtain a pet disease diagnosis semantic feature vector; a pet diagnosis and treatment result generation module 150, which is used to obtain the diagnosis result of the pet's disease based on the pet disease diagnosis semantic feature vector.

[0030] Specifically, in the technical solution of the present application, firstly, a text description of pet symptoms and information on the pet's age, breed, and medical history are obtained. It should be understood that the text description of pet symptoms contains information on abnormal performance of pets such as loss of appetite, vomiting, diarrhea, and coughing, as well as information on the duration and frequency of pet symptoms and behavioral changes of pets. The text description of pet symptoms is the most direct and critical source of information in the diagnosis process. By analyzing the detailed description of the symptoms, the possible range of diseases can be narrowed, providing strong support for subsequent in-depth analysis and final diagnosis. The age, breed, and medical history information of the pet provide important background information for diagnosis, helping the system to understand the health status of the pet more comprehensively. Specifically, the age of the pet has an important influence on the incidence and manifestation of certain diseases. Pets of different breeds also have their specific genetic characteristics and susceptibility to diseases. Medical history information also has a certain reference value for understanding the current symptoms. For example, if the pet has a history of allergies before, the current symptoms may be related to allergic reactions. In general, by comprehensively considering the text description of pet symptoms and the age, breed, and medical history information of the pet, potential disease characteristics can be more accurately identified, thereby achieving more accurate pet disease diagnosis.

[0031] Figure 2 FIG. 1 is a block diagram of a pet symptom data semantic encoding module in a pet disease intelligent diagnosis and treatment data processing system according to an embodiment of the present application. Figure 2 As shown, the pet symptom manifestation data semantic encoding module 120 includes: a pet symptom manifestation word embedding unit 121, which is used to pass the pet symptom manifestation text description through a word embedding layer after word segmentation to obtain a sequence of pet symptom manifestation related word embedding vectors; a pet symptom manifestation related word semantic extraction unit 122, which is used to pass the sequence of pet symptom manifestation related word embedding vectors through a pet symptom manifestation semantic encoder to obtain multiple pet symptom manifestation related word semantic understanding feature vectors; a pet symptom manifestation semantic association feature generation unit 123, which is used to arrange the multiple pet symptom manifestation related word semantic understanding feature vectors into a one-dimensional feature vector and then pass it through a multi-scale pet symptom manifestation local association semantic understander to obtain the pet symptom manifestation semantic association feature vector.

[0032] The Chinese text of pet symptom text description is composed of a continuous sequence of characters, and there is no obvious boundary between words in the sentence. In order to perform effective text analysis, the continuous text description needs to be segmented to obtain meaningful vocabulary units. After segmentation, the processing unit is reduced to the vocabulary level, which can perform semantic analysis more effectively. That is, the pet symptom text description is first segmented. Considering that in the deep learning model, the word embedding layer can map the segmented vocabulary to a high-dimensional space, so that semantically similar words have a similar distance in the vector space. This representation method helps the model capture the semantic relationship between words, thereby improving the accuracy and efficiency of text processing. The segmented pet symptom text description needs to be passed through the word embedding layer to obtain a sequence of pet symptom-related word embedding vectors. The word embedding layer converts the segmented vocabulary into vector representations. These vectors contain not only the semantic information of the vocabulary, but also the contextual relationship between the vocabulary, which enables the model to have a deeper understanding of the key information such as symptom manifestations, severity, and duration in the text description, thereby more accurately diagnosing pet diseases.

[0033] In an embodiment of the present application, a possible implementation method of passing the pet symptom manifestation text description through a word embedding layer after word segmentation to obtain a sequence of pet symptom manifestation-related word embedding vectors may be: performing word segmentation processing on the pet symptom manifestation text description to obtain a plurality of pet symptom manifestation-related words; passing the plurality of pet symptom manifestation-related words through the word embedding layer to convert each of the plurality of pet symptom manifestation-related words into a pet symptom manifestation-related word embedding vector to obtain a sequence of pet symptom manifestation-related word embedding vectors, wherein the word embedding layer uses a learnable embedding matrix to embed code each of the pet symptom manifestation-related words.

[0034] In order to deeply explore the semantic information in the word embedding vector and extract the key semantic features related to the pet symptom manifestation, these features not only contain the description of the symptoms themselves, but also imply the correlation and importance information between the symptoms. It is necessary to pass the sequence of the pet symptom manifestation related word embedding vectors through the pet symptom manifestation semantic encoder to obtain multiple pet symptom manifestation related word semantic understanding feature vectors. In this application, the pet symptom manifestation semantic encoder is a contextual semantic encoder based on a converter. The converter is a deep learning model based on the self-attention mechanism. It completely abandons the structure of traditional recurrent neural networks and convolutional neural networks and relies entirely on the self-attention mechanism to process sequence data. Traditional recurrent neural network models are prone to gradient vanishing or gradient exploding problems when processing long sequences, while the converter can directly calculate the dependency between any two positions in the sequence through the self-attention mechanism, thereby effectively capturing long-distance dependencies. Through the self-attention mechanism of the converter, the semantic encoder can capture the interactions and relationships between the various words in the symptom description, thereby enhancing the model's overall understanding of the text content and helping to improve the accuracy and reliability of disease diagnosis.

[0035] In order to achieve the integration of multiple key information in pet symptom manifestations, which include but are not limited to the type, severity, duration, etc. of symptoms, it is necessary to arrange the semantic understanding feature vectors of the multiple pet symptom manifestation related words into a one-dimensional feature vector. In pet symptom manifestations, the importance and influence of different symptoms may vary. Some symptoms may be more significant and critical, while some symptoms are relatively minor. In order to capture the feature information of different scales, so as to more comprehensively reflect the overall situation of pet symptom manifestations, it is necessary to perform multi-scale processing on the arranged one-dimensional feature vectors. That is, after arranging the semantic understanding feature vectors of the multiple pet symptom manifestation related words into a one-dimensional feature vector, a multi-scale pet symptom manifestation local association semantic understander is used to obtain the pet symptom manifestation semantic association feature vector. The multi-scale pet symptom manifestation local association semantic understander described in the present application is a convolutional neural network model comprising a first convolutional layer and a second convolutional layer, wherein the first convolutional layer uses a one-dimensional convolutional kernel of a first scale, and the second convolutional layer uses a one-dimensional convolutional kernel of a second scale, and the first scale is different from the second scale. The multi-scale pet symptom local correlation semantic understander can improve the generalization ability of the model by learning feature information at different scales, enabling it to better adapt to the manifestations of different pets and different symptoms.

[0036] In an embodiment of the present application, after arranging the multiple pet symptom manifestation related word semantic understanding feature vectors into a one-dimensional feature vector, a possible implementation method of obtaining a pet symptom manifestation semantic association feature vector through a multi-scale pet symptom manifestation local association semantic understander may be: inputting the one-dimensional feature vector into the first convolution layer of the multi-scale pet symptom manifestation local association semantic understander to obtain a first-scale pet symptom manifestation semantic association feature vector, wherein the first convolution layer has a one-dimensional convolution kernel of a first scale; inputting the one-dimensional feature vector into the second convolution layer of the multi-scale pet symptom manifestation local association semantic understander to obtain a second-scale pet symptom manifestation semantic association feature vector, wherein the second convolution layer has a one-dimensional convolution kernel of a second scale, and the first scale is different from the second scale; cascading the first-scale pet symptom manifestation semantic association feature vector and the second-scale pet symptom manifestation semantic association feature vector to obtain the pet symptom manifestation semantic association feature vector.

[0037] Figure 3 FIG. 1 is a block diagram of a semantic coding module for pet disease auxiliary information in a pet disease intelligent diagnosis and treatment data processing system according to an embodiment of the present application. Figure 3 As shown, the pet disease auxiliary information semantic encoding module 130 includes: a pet disease auxiliary information preprocessing unit 131, which is used to preprocess the pet age, breed, and previous medical history information to obtain preprocessed pet disease diagnosis auxiliary reference information; a pet disease diagnosis auxiliary semantic encoding feature extraction unit 132, which is used to segment the preprocessed pet disease diagnosis auxiliary reference information and pass it through a pet disease diagnosis auxiliary information semantic encoder to obtain the pet disease diagnosis auxiliary semantic encoding feature vector.

[0038] The pet's age, breed, and medical history information are important auxiliary references for disease diagnosis. Their accuracy and completeness directly affect the reliability of the diagnosis results. In order to convert the original pet's age, breed, and medical history information into a form suitable for subsequent analysis and processing to improve the quality and availability of the data, it is necessary to preprocess the pet's age, breed, and medical history information to obtain preprocessed auxiliary reference information for pet disease diagnosis. During the preprocessing process, some problems with the original data can be identified and corrected. For example, for breed information, a unified naming standard can be used to avoid confusion caused by spelling errors or aliases; for age information, it can be uniformly converted into textual expressions.

[0039] Next, the pre-processed pet disease diagnosis auxiliary reference information is segmented and then passed through a pet disease diagnosis auxiliary information semantic encoder to obtain a pet disease diagnosis auxiliary semantic encoding feature vector. The pet disease diagnosis auxiliary information semantic encoder described in the present application is a bidirectional long short-term memory neural network model including an embedding layer. It should be understood that in order to facilitate the extraction of semantic features from the pre-processed pet disease diagnosis auxiliary reference information, the pre-processed pet disease diagnosis auxiliary reference information can first be segmented. The word segmentation can convert the complex reference information text into word units, thereby obtaining a reference auxiliary information key vocabulary sequence. The embedding layer can then convert the reference auxiliary information key vocabulary sequence into a high-dimensional numerical vector to further capture the semantic information of the reference auxiliary information keywords. The bidirectional long short-term memory neural network model can simultaneously consider the contextual information of the input vector sequence, so as to better understand the complex context and semantic relationships in the pet disease diagnosis auxiliary reference information.

[0040] The pet symptom expression semantic association feature vector mainly captures the symptoms currently manifested by the pet and the correlation between them. This information is the direct basis for disease diagnosis; the pet disease diagnosis auxiliary semantic coding feature vector contains auxiliary information such as the pet's age, breed, and past medical history. Although this information does not directly reflect the current symptoms, it plays an important role in determining the type of disease; in order to obtain a comprehensive view that includes both direct symptoms and background information, which helps to diagnose the disease more accurately, in the technical solution of the present application, the pet symptom expression semantic association feature vector and the pet disease diagnosis auxiliary semantic coding feature vector are fused to obtain the pet disease diagnosis semantic feature vector. A single source of information may have limitations. For example, some symptoms may appear in different diseases, and auxiliary information can help distinguish these diseases. That is, the fused vector takes into account information from multiple aspects, reduces the possibility of misdiagnosis, and thus improves the accuracy of diagnosis.

[0041] The pet diagnosis and treatment result generation module includes: a pet disease diagnosis semantic feature optimization unit, which is used to perform feature offset correction based on backpropagation on the pet disease diagnosis semantic feature vector to obtain an optimized pet disease diagnosis semantic feature vector; and an optimized pet disease diagnosis semantic feature classification unit, which is used to pass the optimized pet disease diagnosis semantic feature vector through a classifier-based pet disease type result generator to obtain a classification result, wherein the classification result is a label of the disease type suffered by the pet.

[0042] In particular, in the actual text description of pet symptom manifestations and the pet's age, breed, and medical history information, there may be incomplete sentences or some irrelevant information. These data will cause noise in the input text data, and in the process of semantic encoding of the input data, the semantic encoder may not be able to fully understand all the contextual relationships of the input text content, especially when dealing with complex sentence structures or polysemous words, which may lead to inaccurate semantic local features. In addition, when fusing the semantic features of pet symptom manifestations with the auxiliary semantic features of pet disease diagnosis, the final pet disease diagnosis semantic feature vector may have local feature disturbances due to improper weight distribution or information asynchrony. In order to ensure that the classifier can accurately identify the pet's disease even in the presence of these interferences, therefore, in the technical solution of the present application, it is necessary to perform a feature offset correction based on backpropagation on the pet disease diagnosis semantic feature vector to obtain an optimized pet disease diagnosis semantic feature vector.

[0043] Specifically, Figure 4 FIG. 1 is a block diagram of a pet disease diagnosis semantic feature optimization unit in a pet disease intelligent diagnosis and treatment data processing system according to an embodiment of the present application. Figure 4 As shown, the pet disease diagnosis semantic feature optimization unit 151 includes: a classification weight matrix determination subunit 1511, which is used to determine the first classification weight matrix and the second classification weight matrix of the classifier before and after each iterative update; a pet disease diagnosis modulation feature generation subunit 1512, which is used to matrix multiply the first classification weight matrix and the second classification weight matrix with the pet disease diagnosis semantic feature vector respectively to obtain a first pet disease diagnosis modulation feature vector and a second pet disease diagnosis modulation feature vector; a pet disease diagnosis shift information representation subunit 1513, which is used to calculate the position difference between the first pet disease diagnosis modulation feature vector and the second pet disease diagnosis modulation feature vector to obtain a pet disease diagnosis shift information representation vector; a shift compensation scaling factor calculation subunit 1514, which is used to Calculate the F norm of the positional mean vector between the first pet disease diagnosis modulated feature vector and the second pet disease diagnosis modulated feature vector as a shift compensation scaling factor; a pet diagnosis backpropagation shift compensation representation generating subunit 1515 is used to linearly scale the pet disease diagnosis shift information representation vector with the shift compensation scaling factor to obtain a scaled pet disease diagnosis shift information representation vector, and input the scaled pet disease diagnosis shift information representation vector into a Sigmoid activation function to obtain a pet diagnosis backpropagation shift compensation representation vector; a pet disease diagnosis semantic optimization feature generating subunit 1516 is used to calculate the positional point multiplication of the pet diagnosis backpropagation shift compensation representation vector and the pet disease diagnosis semantic feature vector to obtain the optimized pet disease diagnosis semantic feature vector.

[0044] In an embodiment of the present application, another expression of performing back-propagation-based feature offset correction on the pet disease diagnosis semantic feature vector to obtain an optimized pet disease diagnosis semantic feature vector may be: processing the pet disease diagnosis semantic feature vector with the following formula to obtain an optimized pet disease diagnosis semantic feature vector; wherein the formula is:

[0045]

[0046] Wherein, M1 is the first classification weight matrix, M2 is the second classification weight matrix, V is the pet disease diagnosis semantic feature vector, represents matrix multiplication, Indicates adding by position point, ‖·‖ F represents the F norm of the feature vector, S represents the shift compensation scaling factor, ⊙ represents the point multiplication by position, represents point-by-point subtraction, sigmoid represents an activation function, and V′ represents the optimized pet disease diagnosis semantic feature vector.

[0047] More specifically, in an embodiment of the present application, the pet diagnosis backward propagation shift compensation representation generation subunit is used to: create a Spring Boot project; create a service class in the Spring Boot project, and the service class is used to handle the logic of linear scaling and the Sigmoid activation function; create a controller class in the Spring Boot project, and the controller class is used to process HTTP requests and call the service class; configure the application properties of the Spring Boot project, and package the Spring Boot project and deploy it on the server.

[0048] In the embodiment of the present application, the pet disease diagnosis shift information representation shift information representation vector is linearly scaled by the shift compensation scaling factor to obtain a scaled pet disease diagnosis shift information representation shift information representation vector, and the scaled pet disease diagnosis shift information representation shift information representation vector is input into a Sigmoid activation function to obtain a pet diagnosis shift information representation back-propagation shift compensation representation vector. Part of the code is specifically:

[0049]

[0050]

[0051]

[0052] It should be understood that the advantages of using the Spring web mechanism to deploy and implement "linearly scaling the pet disease diagnosis shift information representation shift information representation vector with the shift compensation scaling factor to obtain the scaled pet disease diagnosis shift information representation shift information representation vector, and inputting the scaled pet disease diagnosis shift information representation shift information representation vector into the Sigmoid activation function to obtain the pet diagnosis shift information representation backpropagation shift compensation representation vector" include: a. Spring's design philosophy encourages dividing applications into multiple layers (such as controllers, services, and data access layers), which makes the code easier to manage and maintain, and the responsibilities of each layer are clear, which facilitates development and testing; b. The Spring framework provides rich functionality and flexibility, and can easily add new features or modify existing features, and can easily add new algorithms or data processing logic without refactoring the entire application; c. Spring supports asynchronous request processing, which can improve system performance in high concurrency scenarios, especially when processing large amounts of data; d. Spring Web provides powerful RESTful API support makes it easy to create and manage HTTP requests, and can easily implement data interaction in JSON format, which is suitable for the needs of modern Web applications; e. The Spring framework has a huge community support and rich documentation, so developers can easily find solutions to problems or best practices.

[0053] In order to enhance the robustness of the classifier to local disturbances in the input feature vector, the pet disease diagnosis semantic feature vector is subjected to shift compensation based on backpropagation representation. The key to the shift compensation based on backpropagation representation of the pet disease diagnosis semantic feature vector is to capture the information of the classification weight matrix changing with the training process, and to use this information to generate a pet disease diagnosis shift information representation vector, and then generate a pet diagnosis backpropagation shift compensation representation vector through scaling and nonlinear transformation (Sigmoid function). In this way, the pet disease diagnosis semantic feature vector is described by the local shift of the feature by supporting the distribution difference of different classification scenarios of the classifier during the training process, so that not only the displacement of the feature vector caused by the change of the classification weight matrix can be quantified, but also a compensation vector can be generated according to the displacement information. The role of the compensation vector is to weight and adjust each element in the feature vector according to the degree of influence of the disturbance, thereby reducing the influence of the disturbance on the classification result. In this way, the classifier's robustness to local disturbances in the input data can be significantly improved, the classifier's stability and generalization ability can be enhanced, and by reducing the impact of local disturbances on the classification results, the classifier can maintain a high classification accuracy when facing noisy or variant data.

[0054] Finally, the optimized pet disease diagnosis semantic feature vector is passed through a pet disease type result generator based on a classifier to obtain a classification result, and the classification result is a label of the disease type suffered by the pet. It should be understood that a classifier based on a large amount of training data can learn the complex relationship between the disease and the symptoms, so as to output an accurate disease type label given a feature vector. That is, the classifier can automatically extract key information from a complex feature vector and map it to a predefined disease type label. The classification result directly gives the specific type label of the disease suffered by the pet, providing a clear diagnosis basis for the pet owner, which helps them to formulate targeted treatment plans and nursing measures. Obtaining the pet disease type result through the classifier, this automated processing process greatly improves the efficiency of disease diagnosis, reduces the need for manual intervention, makes large-scale pet disease diagnosis possible, and by automatically generating classification results, the intelligent diagnosis and treatment system can significantly shorten the diagnosis time and improve the efficiency of diagnosis and treatment.

[0055] In summary, the pet disease intelligent diagnosis and treatment data processing system 100 based on the embodiment of the present application is explained, which adopts the semantic analysis technology based on deep learning, obtains the semantic features of pet symptom manifestations by semantically encoding the text description of pet symptom manifestations, and obtains the semantic features of pet diagnosis auxiliary information by semantically encoding the pet age, breed, and past medical history information, and comprehensively considers the semantic features of pet symptom manifestations and the semantic features of pet diagnosis auxiliary information, so as to realize the diagnosis of pet diseases. In this way, the accuracy of pet disease diagnosis can be improved, and timely pet disease diagnosis and treatment assistance can be provided during non-working hours.

[0056] As described above, the pet disease intelligent diagnosis and treatment data processing system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server for pet disease intelligent diagnosis and treatment data processing. In one example, the pet disease intelligent diagnosis and treatment data processing system 100 according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the pet disease intelligent diagnosis and treatment data processing system 100 can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the pet disease intelligent diagnosis and treatment data processing system 100 can also be one of the many hardware modules of the terminal device.

[0057] Alternatively, in another example, the pet disease intelligent diagnosis and treatment data processing system 100 and the terminal device may also be separate devices, and the pet disease intelligent diagnosis and treatment data processing system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0058] Figure 5 FIG. 1 is a flow chart of a method for processing pet disease intelligent diagnosis and treatment data according to an embodiment of the present application. Figure 5 As shown, in the method for processing data for intelligent diagnosis and treatment of pet diseases, the following are included: S110, obtaining a text description of pet symptom manifestations and the pet's age, breed, and past medical history information; S120, semantically encoding the text description of pet symptom manifestations to obtain a pet symptom manifestation semantic association feature vector; S130, semantically encoding the pet's age, breed, and past medical history information to obtain a pet disease diagnosis auxiliary semantic coding feature vector; S140, fusing the pet symptom manifestation semantic association feature vector and the pet disease diagnosis auxiliary semantic coding feature vector to obtain a pet disease diagnosis semantic feature vector; S150, obtaining a diagnosis result of the pet's disease based on the pet disease diagnosis semantic feature vector.

[0059] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned pet disease intelligent diagnosis and treatment data processing method have been referred to above. Figures 1 to 4 The invention has been introduced in detail in the description of the intelligent diagnosis and treatment data processing system for pet diseases, and therefore, its repeated description will be omitted.

[0060] In summary, the method for processing data for intelligent diagnosis and treatment of pet diseases based on the embodiment of the present application is explained, which adopts semantic analysis technology based on deep learning, obtains semantic features of pet symptom manifestations by semantically encoding the text description of pet symptom manifestations, and obtains semantic features of pet diagnosis auxiliary information by semantically encoding the pet age, breed, and past medical history information, and comprehensively considers the semantic features of pet symptom manifestations and the semantic features of pet diagnosis auxiliary information, thereby realizing the diagnosis of pet diseases. In this way, the accuracy of pet disease diagnosis can be improved, and timely pet disease diagnosis and treatment assistance can be provided during non-working hours.

Claims

1. A pet disease intelligent diagnosis and treatment data processing system, characterized in that: include: The pet diagnosis and treatment related data collection module is used to obtain text descriptions of pet symptoms and information about the pet's age, breed, and medical history; A pet symptom data semantic encoding module, used for semantically encoding the pet symptom text description to obtain a pet symptom semantic association feature vector; A pet disease auxiliary information semantic coding module is used to semantically encode the pet's age, breed, and past medical history information to obtain a pet disease diagnosis auxiliary semantic coding feature vector; A semantic coding feature fusion module, used for fusing the pet symptom manifestation semantic association feature vector and the pet disease diagnosis auxiliary semantic coding feature vector to obtain a pet disease diagnosis semantic feature vector; The pet diagnosis and treatment result generating module is used to obtain the diagnosis result of the pet's disease based on the pet disease diagnosis semantic feature vector.

2. The pet disease intelligent diagnosis and treatment data processing system according to claim 1 is characterized in that: The pet symptom data semantic encoding module includes: A pet symptom manifestation word embedding unit, used to pass the pet symptom manifestation text description through a word embedding layer after word segmentation to obtain a sequence of pet symptom manifestation related word embedding vectors; A pet symptom manifestation related word semantic extraction unit, used for passing the sequence of pet symptom manifestation related word embedding vectors through a pet symptom manifestation semantic encoder to obtain a plurality of pet symptom manifestation related word semantic understanding feature vectors; A pet symptom manifestation semantic association feature generation unit is used to arrange the multiple pet symptom manifestation related word semantic understanding feature vectors into a one-dimensional feature vector and then pass it through a multi-scale pet symptom manifestation local association semantic understander to obtain the pet symptom manifestation semantic association feature vector.

3. The pet disease intelligent diagnosis and treatment data processing system according to claim 2 is characterized in that: The pet disease auxiliary information semantic encoding module includes: A pet disease auxiliary information preprocessing unit, used to preprocess the pet's age, breed, and past medical history information to obtain preprocessed pet disease diagnosis auxiliary reference information; The pet disease diagnosis auxiliary semantic coding feature extraction unit is used to segment the preprocessed pet disease diagnosis auxiliary reference information and pass it through a pet disease diagnosis auxiliary information semantic encoder to obtain the pet disease diagnosis auxiliary semantic coding feature vector.

4. The pet disease intelligent diagnosis and treatment data processing system according to claim 3 is characterized in that: The pet symptom manifestation semantic encoder is a converter-based contextual semantic encoder, the pet disease diagnosis auxiliary information semantic encoder is a bidirectional long short-term memory neural network model including an embedding layer, and the multi-scale pet symptom manifestation local association semantic understander is a convolutional neural network model including a first convolutional layer and a second convolutional layer, wherein the first convolutional layer uses a one-dimensional convolutional kernel of a first scale, and the second convolutional layer uses a one-dimensional convolutional kernel of a second scale, and the first scale is different from the second scale.

5. The pet disease intelligent diagnosis and treatment data processing system according to claim 4 is characterized in that: The pet diagnosis and treatment result generating module includes: A pet disease diagnosis semantic feature optimization unit, used for performing a feature offset correction based on back propagation on the pet disease diagnosis semantic feature vector to obtain an optimized pet disease diagnosis semantic feature vector; The optimized pet disease diagnosis semantic feature classification unit is used to pass the optimized pet disease diagnosis semantic feature vector through a classifier-based pet disease type result generator to obtain a classification result, wherein the classification result is a label of the disease type suffered by the pet.

6. The pet disease intelligent diagnosis and treatment data processing system according to claim 5 is characterized in that: The pet disease diagnosis semantic feature optimization unit comprises: A classification weight matrix determination subunit, used to determine a first classification weight matrix and a second classification weight matrix of the classifier before and after each iterative update; A pet disease diagnosis modulation feature generation subunit, used for matrix multiplying the first classification weight matrix and the second classification weight matrix with the pet disease diagnosis semantic feature vector respectively to obtain a first pet disease diagnosis modulation feature vector and a second pet disease diagnosis modulation feature vector; a pet disease diagnosis shift information representation subunit, configured to calculate a positional difference between the first pet disease diagnosis modulation feature vector and the second pet disease diagnosis modulation feature vector to obtain a pet disease diagnosis shift information representation vector; a shift compensation scaling factor calculation subunit, used to calculate the F norm of the position mean vector between the first pet disease diagnosis modulation feature vector and the second pet disease diagnosis modulation feature vector as a shift compensation scaling factor; A pet diagnosis backpropagation shift compensation representation generating subunit is used to linearly scale the pet disease diagnosis shift information representation vector with the shift compensation scaling factor to obtain a scaled pet disease diagnosis shift information representation vector, and input the scaled pet disease diagnosis shift information representation vector into a Sigmoid activation function to obtain a pet diagnosis backpropagation shift compensation representation vector; The pet disease diagnosis semantic optimization feature generation subunit is used to calculate the point-by-point multiplication of the pet diagnosis back-propagation shift compensation representation vector and the pet disease diagnosis semantic feature vector to obtain the optimized pet disease diagnosis semantic feature vector.

7. The pet disease intelligent diagnosis and treatment data processing system according to claim 6, characterized in that: The pet diagnosis back-propagation shift compensation representation generation subunit is used to: Create a Spring Boot project. Create a service class in the Spring Boot project, which is used to process the logic of linear scaling and the Sigmoid activation function; Create a controller class in the Spring Boot project, which is used to process HTTP requests and call the service class; Configure the application properties of the Spring Boot project, and package the Spring Boot project and deploy it on the server.

8. A method for processing pet disease intelligent diagnosis and treatment data, characterized in that: include: Obtain text descriptions of pet symptoms and information about the pet's age, breed, and medical history; Performing semantic encoding on the pet symptom manifestation text description to obtain a pet symptom manifestation semantic association feature vector; Semantically encoding the pet's age, breed, and past medical history information to obtain a pet disease diagnosis-aided semantic encoding feature vector; Fusion of the pet symptom manifestation semantic association feature vector and the pet disease diagnosis auxiliary semantic coding feature vector to obtain a pet disease diagnosis semantic feature vector; Based on the pet disease diagnosis semantic feature vector, a diagnosis result of the pet disease is obtained.

9. The method for processing pet disease intelligent diagnosis and treatment data according to claim 8, characterized in that: Semantically encoding the pet symptom text description to obtain a pet symptom semantic association feature vector includes: The pet symptom text description is passed through a word embedding layer after word segmentation to obtain a sequence of pet symptom related word embedding vectors; Passing the sequence of pet symptom manifestation related word embedding vectors through a pet symptom manifestation semantic encoder to obtain multiple pet symptom manifestation related word semantic understanding feature vectors; The plurality of pet symptom manifestation related word semantic understanding feature vectors are arranged into a one-dimensional feature vector and then passed through a multi-scale pet symptom manifestation local association semantic understander to obtain the pet symptom manifestation semantic association feature vector.

10. The method for processing pet disease intelligent diagnosis and treatment data according to claim 9, characterized in that: Based on the pet disease diagnosis semantic feature vector, the diagnosis result of the pet disease is obtained, including: Performing feature offset correction based on back-propagation on the pet disease diagnosis semantic feature vector to obtain an optimized pet disease diagnosis semantic feature vector; The optimized pet disease diagnosis semantic feature vector is passed through a classifier-based pet disease type result generator to obtain a classification result, and the classification result is a label of the disease type suffered by the pet.

Images