Medical image diagnosis management method and system based on Internet

Abstract

The invention relates to the technical field of medical image diagnosis management, in particular to an internet-based medical image diagnosis management method and system, and the method comprises the following steps: obtaining initial medical image information of a patient through an internet medical image diagnosis management system; performing information adjustment on the initial medical image information to obtain processed medical image information; constructing an image correction model based on the medical image information, and analyzing and adjusting the medical image of the patient by using the image correction model to obtain a standardized medical image; setting a medical image feature fusion mechanism according to the medical image information, and obtaining medical image feature information according to the medical image feature fusion mechanism; and providing medical diagnosis service for the patient in combination with the standardized medical image, the medical image feature information and the medical image information. The mathematical model is used for processing, analyzing and verifying the medical image, so that the accuracy of a medical diagnosis result is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image diagnosis management, and in particular to an Internet-based medical image diagnosis management method and system. Background Art

[0002] Existing medical imaging systems collect a large amount of medical imaging information. How to efficiently integrate and fully utilize medical imaging resources to accelerate and optimize the doctor's diagnosis process is a core issue that needs to be urgently addressed in Internet medical imaging diagnosis management technology. Traditional systems are subject to limitations such as lagging knowledge updates, insufficient system robustness, a single reasoning mechanism, and susceptibility to subjective influences. The efficiency and accuracy of existing systems when processing complex and changeable medical images are affected.

[0003] Since medical imaging information has the characteristics of large data volume, close multi-dimensional correlation and highly specialized objects, its data mining and analysis process is far more complicated than conventional databases. At present, when processing complex data in specific medical fields, there is a lack of efficient and targeted data mining and analysis tools, which in turn restricts the effective use of medical imaging information.

[0004] Therefore, it is necessary to explore and develop new methods for medical image information management. Not only can we deeply integrate advanced medical image processing technology with data mining technology, but we also need to achieve efficient extraction, accurate loading and intelligent classification of medical image features, and mine valuable association rules and deep information from massive image data. This will enable the analysis of medical image information, and combine Internet technology to build a more comprehensive and intuitive diagnosis and treatment assistance platform, promote the advancement of medical image diagnosis management technology, and enhance the intelligence level of medical diagnosis systems. Summary of the invention

[0005] In view of the shortcomings of existing methods and the needs of practical applications, the medical image processing and analysis technology are integrated to accurately extract and dynamically load the relevant features of medical images, and to achieve efficient organization and management of medical image information through intelligent classification algorithms. Combining Internet technology to build a comprehensive, intuitive and objective diagnosis and treatment system can not only significantly improve the accuracy and work efficiency of medical image diagnosis, but also contribute to the development of medical image management technology, promote the medical diagnosis system to a higher level of intelligence, digitization and networking, and provide patients with a more convenient, accurate and personalized medical service experience. On the one hand, the present invention provides an Internet-based medical image diagnosis method, which includes the following steps: obtaining the initial medical image information of the patient through an Internet medical image diagnosis management system; adjusting the information of the initial medical image information to obtain the processed medical image information; constructing an image correction model based on the medical image information, and using the image correction model to analyze and adjust the patient's medical image to obtain a standardized medical image; setting a medical image feature fusion mechanism according to the medical image information, and obtaining medical image feature information according to the medical image feature fusion mechanism; combining the standardized medical image, the medical image feature information and the medical image information to provide medical diagnosis services for the patient. The present invention effectively manages image resources through the medical image management system, automatically processes and analyzes medical image information, reduces manual interpretation time, and improves the accuracy of diagnosis, which helps to achieve personalized and targeted medical diagnosis services.

[0006] Optionally, the information adjustment of the initial medical image information to obtain the processed medical image information includes: configuring a cloud image receiving and forwarding module in the Internet medical image diagnosis and management system; setting an artificial neural network model in the cloud image receiving and forwarding module, and using the artificial neural network model to adjust the information of the initial medical image information to obtain the processed medical image information. The artificial neural network model of the present invention can identify, analyze and adjust medical image information without human intervention, greatly improving the speed and efficiency of information processing, and can more accurately identify subtle changes in medical images, reducing the occurrence of manual misdiagnosis and missed diagnosis.

[0007] Optionally, the processed medical image information satisfies the following relationship:

[0008]

[0009] Among them, PMI represents the processed medical image information, R represents the artificial neural network activation model, m represents the total number of neurons in the artificial neural network, and I i represents the medical image information input by the i-th neuron, Indicates the connection weight of the ith neuron in the artificial neural network, and τ indicates the bias value of the artificial neural network. In the model of the present invention, the connection weight and bias value are key parameters in the learning process of the artificial neural network. Automatic adjustment of the parameters can adapt to different medical imaging data, so that the system can continuously improve the accuracy and robustness of diagnosis.

[0010] Optionally, the setting of the medical image feature fusion mechanism according to the medical image information includes: configuring an image processing and storage module in the Internet medical image diagnosis and management system; the image processing and storage module sets the medical image feature fusion mechanism according to the processed medical image information, and the medical image feature fusion mechanism includes a neural network branch feature information model and an image feature fusion model. The present invention sets the medical image feature fusion mechanism according to the processed medical image information, and configures the image processing and storage module to support the above process, which can enhance the feature extraction capability, optimize the feature fusion effect, and promote the intelligent development of medical imaging.

[0011] Optionally, the neural network branch feature information model satisfies the following relationship:

[0012]

[0013] Among them, n It represents the feature information set of different branches in the artificial neural network. L represents the first feature size of the nth branch, Y represents the second feature size of the nth branch, m represents the total number of nodes in the nth branch, and PMI m (l, y) represents the feature information of the mth node in the branch, λ m Indicates the weight corresponding to the mth node in the branch. The neural network branch feature information model of the present invention can extract feature information of different branches, so that the system can more comprehensively capture key information in medical images and improve the accuracy of diagnosis.

[0014] Optionally, the image feature fusion model satisfies the following relationship:

[0015]

[0016] Among them, FFI represents the final output of medical image feature fusion information, fix represents the fusion operation function of medical image features, N represents the number of branches of medical images in the artificial neural network, g represents the activation function of image feature information, and ffi iRepresents the feature information set of the i-th branch in the artificial neural network. The image feature fusion model of the present invention fuses the feature information extracted from different branches, so that the system can integrate information from multiple aspects to form a more comprehensive and accurate medical image feature representation, which helps doctors to more accurately judge the condition and formulate personalized treatment plans.

[0017] Optionally, the obtaining of medical image feature information according to the medical image feature fusion mechanism includes: obtaining a feature information set of different branches in the artificial neural network through the neural network branch feature information model; and obtaining fusion information of medical image features based on the feature information set of different branches and the image feature fusion model. The process of obtaining medical image feature information according to the medical image feature fusion mechanism of the present invention can improve the comprehensiveness and accuracy of feature extraction, enhance the complementarity of feature information, and improve the robustness and reliability of diagnosis.

[0018] Optionally, the image correction model is constructed based on the medical image information, and the image correction model is used to analyze and adjust the patient's medical image to obtain a standardized medical image, including: configuring an image intelligent diagnosis module in the Internet medical image diagnosis management system; constructing an image correction model in the image intelligent diagnosis module based on the medical image information; and using the image correction model to analyze and adjust the patient's medical image to obtain a standardized medical image. The image correction model of the present invention analyzes and adjusts the patient's medical image, which can reduce image quality problems caused by equipment differences, acquisition conditions, and other factors, thereby ensuring that doctors have high-quality image data when making diagnoses, which helps to improve the accuracy of diagnosis and reduce the occurrence of misdiagnosis and missed diagnosis.

[0019] Optionally, the image correction model satisfies the following relationship:

[0020]

[0021] Among them, J C represents the gray value of the image after standardization, σ represents the comprehensive influence coefficient of abnormal noise in the image, J represents the initial gray value of the image, and J min Indicates the minimum gray value of the image, J max Indicates the maximum grayscale value of an image. The image correction model of the present invention can reduce noise interference, optimize grayscale distribution, enhance image contrast, improve image visualization, support quantitative analysis, and promote standardization and normalization of medical images.

[0022] In the second aspect, in order to efficiently execute the Internet-based medical imaging diagnosis method provided by the present invention, the present invention also provides an Internet-based medical imaging diagnosis system, the system comprising a medical imaging acquisition module, a cloud-based imaging receiving and forwarding module, an imaging processing and storage module, and an imaging intelligent diagnosis module, the medical imaging acquisition module, the cloud-based imaging receiving and forwarding module, the imaging processing and storage module, and the imaging intelligent diagnosis module are interconnected, the system executes the Internet-based medical imaging diagnosis method as described in the first aspect of the present invention, the system has a compact structure and stable performance, and can stably execute the Internet-based medical imaging diagnosis method provided by the present invention, thereby improving the overall applicability and practical application capabilities of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flow chart of the Internet-based medical imaging diagnosis method of the present invention;

[0024] Figure 2 A schematic diagram of the structure of an artificial neural network model of the Internet-based medical imaging diagnosis method of the present invention;

[0025] Figure 3 A schematic diagram of the structure of neurons in the artificial neural network model of the Internet-based medical imaging diagnosis method of the present invention;

[0026] Figure 4 This is a structural diagram of the Internet-based medical imaging diagnosis system of the present invention. DETAILED DESCRIPTION

[0027] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are only for illustration and are not intended to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is obvious to those of ordinary skill in the art that these specific details do not need to be adopted to implement the present invention. In other examples, in order to avoid confusing the present invention, known circuits, software or methods are not specifically described.

[0028] Throughout the specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment of the present invention. Therefore, the phrases "in one embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily all refer to the same embodiment or example. In addition, particular features, structures, or characteristics may be combined in one or more embodiments or examples in any suitable combination and / or subcombination. In addition, it should be understood by those of ordinary skill in the art that the figures provided herein are for illustrative purposes and that the figures are not necessarily drawn to scale.

[0029] The present invention processes, analyzes and adjusts medical images, accurately captures and processes key feature information in medical images in real time, and uses advanced intelligent classification and recognition algorithms to integrate and intelligently manage massive amounts of medical image information. The invention constructs a comprehensive, intuitive and objective medical image diagnosis management method and system, which not only improves the accuracy and timeliness of medical image diagnosis, but also leads the innovation of medical image management technology and promotes the field of medical diagnosis to an advanced stage of intelligence, digitization and networking. This achieves the personalization and precision of medical services and significantly improves the patient's medical experience.

[0030] See also Figure 1 The present invention provides a medical imaging diagnosis method based on the Internet. The implementation steps and specific process of the medical imaging diagnosis method based on the Internet are as follows:

[0031] S1. Obtain the patient's initial medical imaging information through the Internet medical imaging diagnosis management system. The specific steps and implementation contents are as follows:

[0032] The embodiment configures a medical image acquisition module in the Internet medical image diagnosis and management system, and uses the medical image acquisition module to acquire initial medical image information of the patient.

[0033] The hardware in the medical image acquisition module is interconnected. The medical image acquisition module can be seamlessly connected with various digital imaging devices. It includes but is not limited to high-end digital imaging equipment such as DR (digital X-ray imaging system), CT (computed tomography), MRI (magnetic resonance imaging), etc. It can also be closely connected with the Internet medical imaging diagnosis management system through the DICOM (Digital Imaging and Communications in Medicine) protocol, which can ensure the rapid transmission and sharing of medical imaging information. On the other hand, for non-digital medical imaging information, such as traditional films, ultrasound image prints, etc., a high-performance scanner is configured in the medical image acquisition module to realize the conversion of non-digital data, so that the acquisition module can fully adapt to the digital medical system and improve the efficiency and accuracy of diagnosis and treatment.

[0034] Software integration in the medical image acquisition module. At the software level, the medical image acquisition module integrates DICOM receiver technology, which is conducive to establishing a stable and efficient medical image information acquisition module. It can not only capture massive image data from different imaging devices in real time, but also has built-in compression algorithms and encryption mechanisms to ensure the speed and security of image data transmission. The encrypted and compressed medical image information will be directly transmitted to the cloud image receiving and forwarding module, which not only ensures the integrity of medical image information, but also effectively reduces the risk of information leakage.

[0035] The medical image acquisition module has cloud storage and efficient retrieval functions. In order to cope with the growing demand for medical image data, the embodiment adopts advanced cloud storage technology to provide safe and reliable storage space for massive medical image information, which not only improves the flexibility of information access, but also facilitates data backup and recovery. On this basis, an intelligent data indexing device is also configured to allow doctors to quickly locate and retrieve the required medical image information based on various key information such as patient ID, examination time, diagnosis type, etc., shortening the time for doctors to find medical images and providing strong support for clinical decision-making.

[0036] Furthermore, the method for collecting the initial medical imaging information of the patient in the embodiment is only an optional condition of the present invention. In one or some other embodiments, the information acquisition method can be adjusted according to the medical diagnosis needs and medical imaging acquisition requirements, and customized and optimized according to actual needs, thereby improving the flexibility and adaptability of the system, enabling it to be more widely used in various clinical scenarios, and helping doctors to formulate the most appropriate diagnostic strategy according to specific circumstances.

[0037] S2. Adjusting the initial medical image information to obtain processed medical image information. The specific steps and implementation contents are as follows:

[0038] First, a cloud image receiving and forwarding module is configured in the Internet medical image diagnosis management system, and an artificial neural network model is set in the cloud image receiving and forwarding module. At the same time, the artificial neural network model is used to adjust the medical image information to obtain the processed medical image information. The specific contents are as follows:

[0039] The artificial neural network in the embodiment is a mathematical model inspired by bionics, and its design inspiration comes directly from the complex interaction between neurons in the biological nervous system. In the artificial neural network model, each layer or neuron communicates with each other through dense connections, forming a highly flexible information processing framework.

[0040] That is, the artificial neural network model is mainly composed of three main parts: input layer, hidden layer (can have one or more layers) and output layer. The input layer is responsible for receiving external data, that is, the medical image information transmitted from the medical image acquisition module, and then undergoes complex nonlinear transformations through neurons in the hidden layer. The hidden layer is the core of the artificial neural network model, which contains different levels or a series of neurons. Different levels or neurons process the input medical image information through weights and may pass the processed signal to the next layer. Finally, the output layer generates the response or analysis result of the artificial neural network model, that is, obtains the processed medical image information.

[0041] For the basic structure of the artificial neural network model in the embodiment, please refer to Figure 2 , where S1, S2, S3, and Sn represent the medical image information transmitted by different branches, levels, or neurons, and ffi1, ffi2, ffi3, and ffin represent the feature information sets of different branches in the artificial neural network.

[0042] based on Figure 2 It can be seen that the medical image information input during the forward propagation process passes through each layer of the artificial neural network, is calculated layer by layer and passed to the output layer. If the output does not meet expectations, the back propagation algorithm is started, which updates the weights and biases of the neurons in the artificial neural network by calculating the output error and passing this error backward layer by layer. The above process is repeated until the performance of the artificial neural network reaches the predetermined standard or the upper limit of the number of iterations is reached.

[0043] Artificial neural network models can not only simulate the basic functions of biological nervous systems, such as information processing and signal transmission, but also show strong application potential in pattern recognition, classification, prediction and other fields through their unique structure and learning mechanism.

[0044] Furthermore, the artificial neural network model is highly flexible, and the input layer, hidden layer, and output layer work together to learn and express complex mapping relationships between input and output.

[0045] The input layer is responsible for receiving and adjusting the initial medical image information to obtain processed medical image information. The above information is not processed in any way and is directly passed as a signal to the subsequent hidden layer. The number of neurons in the input layer usually matches the feature dimension of the input information to ensure that each feature can be accurately received by the network.

[0046] The hidden layer contains a large number of neurons, which are interconnected in a fully connected manner, and each neuron is equipped with a corresponding function to introduce nonlinear characteristics. The nonlinear characteristics enable the artificial neural network to learn and express complex nonlinear relationships, thereby performing in-depth analysis of medical imaging information. In the embodiment, the hidden layer can have one or more layers, and the increase in the number of layers can improve the complexity and learning ability of the network.

[0047] The output layer is responsible for converting the results of the hidden layer processing into the network's response or prediction to the outside world. The number of neurons in the output layer usually matches the system performance and actual needs. In the task of medical image information classification, the number of neurons in the output layer sets the number and method of categories based on the medical image attributes to obtain the processed medical image information.

[0048] The execution content of a single neuron in the artificial neural network model is as follows:

[0049] In the artificial neural network model, each neuron is an information processing unit. Figure 3 , where I1, I2, I3, and Im represent the medical imaging information input by different neurons, They represent the connection weights corresponding to the nodes in different neurons, τ represents the bias value of the artificial neural network, and PMI represents the processed medical image information.

[0050] Neurons receive multiple medical imaging information or signals from other neurons or external inputs. Before entering the neuron, the information or signal will first be multiplied by the corresponding connection weight to achieve weighted processing of the signal. The purpose of weighted processing is to adjust the contribution of the signal according to the importance or strength of the connection.

[0051] The connection weights of different nodes in an artificial neural network satisfy the following relationship

[0052]

[0053] in, represents the connection weight of the i-th node in the artificial neural network, c i Represents the classification error rate of the i-th node in the artificial neural network.

[0054] The weighted signals are added together to form a total input value, which is then combined with a bias value set inside the neuron, which can fine-tune the total input value to introduce a certain offset, helping the neuron produce richer output results.

[0055] Next, the results of weighted processing and bias value adjustment will be sent to the activation function. The activation function is the core part of the neuron, which determines how the neuron responds to the input signal and can introduce nonlinear characteristics, allowing the neural network to learn and express complex nonlinear relationships. Finally, the activation function calculates the final output of the neuron based on the input value.

[0056] The medical imaging information after the above processing satisfies the following relationship:

[0057]

[0058] Among them, PMI represents the processed medical image information, R represents the artificial neural network activation model, m represents the total number of neurons in the artificial neural network, and I i represents the medical image information input by the i-th neuron, represents the connection weight of the i-th neuron in the artificial neural network, and τ represents the bias value of the artificial neural network.

[0059] In the embodiment, the artificial neural network activation model can be Sigmoid, ReLU, etc., which can introduce nonlinear characteristics so that the neural network can learn and express complex nonlinear relationships.

[0060] Finally, the cloud image receiving and forwarding module will transmit the medical image information processed by the artificial neural network to the image processing and storage module in real time. Cloud processing and forwarding can reduce the computing burden and storage requirements of system equipment, optimize resource allocation, reduce costs, and improve resource utilization efficiency. On the other hand, real-time transmission enables the system to monitor the processing status and transmission of medical image information in real time, promptly discover and deal with possible problems, and doctors can adjust diagnostic strategies based on real-time feedback to improve diagnosis and treatment effects.

[0061] Furthermore, the artificial neural network model in the present embodiment is merely an optional condition of the present embodiment. In one or some other embodiments, the neural network model may be replaced according to medical information requirements and image management system information. Different medical application scenarios and diagnoses require different neural network models to process and analyze data to adapt to different diagnostic scenarios and image management systems, thereby improving the overall flexibility and adaptability of the system.

[0062] S3. Construct an image correction model based on medical image information, and use the image correction model to analyze and adjust the patient's medical images to obtain standardized medical diagnostic images. The specific steps and implementation contents are as follows:

[0063] An image intelligent diagnosis module is configured in the Internet medical image diagnosis management system. An image correction model is constructed in the image intelligent diagnosis module based on medical image information. The image correction model is then used to analyze and adjust the patient's medical images to obtain standardized medical images. The specific implementation contents are as follows:

[0064] In order to overcome the differences in grayscale ranges between different imaging mechanisms and MRI sequences of the same modality, a grayscale normalization strategy is deployed in the embodiment. It is necessary to proportionally adjust the voxel grayscale values ​​of the medical diagnostic images in the Internet medical imaging diagnosis management system to the [0, 1] interval to ensure that the image data in the visualization interaction module is consistent and comparable at the grayscale level, thereby improving the stability and feasibility of the Internet medical imaging diagnosis management method.

[0065] An image correction model is set in the image intelligent diagnosis module. The above image correction model satisfies the following relationship:

[0066]

[0067] Among them, J Crepresents the gray value of the image after standardization, σ represents the comprehensive influence coefficient of abnormal noise in the diagnostic image, J represents the initial gray value of the image, and J min Indicates the minimum gray value of the image, J max Indicates the maximum grayscale value of the image.

[0068] In medical imaging, noise refers to unnecessary, random signal fluctuations in the image. The fluctuations are mainly caused by various factors such as imaging equipment, patient movement, environmental interference, etc. The noise includes but is not limited to quantum noise, electronic noise, display system noise, etc. Each type of noise affects the image quality in different ways and degrees. The indicators for quantifying the comprehensive impact of noise on image quality in the embodiments can more comprehensively and accurately evaluate the impact of noise on the quality of medical images.

[0069] The initial grayscale value of an image refers to the color depth of each pixel in the image before any image processing or correction. The grayscale value is a parameter that describes the color depth of a point in a black and white image, ranging from 0 to 255, where white corresponds to the highest grayscale value, usually 255, and black corresponds to the lowest grayscale value, usually 0. In medical imaging, the grayscale value reflects the density or signal strength of the tissue structure and is an important basis for doctors to diagnose and evaluate diseases.

[0070] In medical image processing and analysis, certain processing and correction of the initial grayscale value can eliminate or reduce the influence of related factors, improve the quality of medical images and the accuracy of diagnosis, and make the comparison and analysis between medical images more reliable and effective.

[0071] In an image, the minimum grayscale value and the maximum grayscale value are two important parameters that describe the grayscale range of the image. The above two parameters are of great significance for understanding the contrast and brightness distribution of the image as well as subsequent image processing and analysis.

[0072] The minimum grayscale value refers to the minimum grayscale value of all pixels in the image, which represents the darkest part of the medical image and usually corresponds to the black or nearly black area. The size of the above minimum grayscale value can reflect the dark details and contrast of the medical image. A smaller minimum grayscale value means that there are darker areas in the image, which helps to show low-density tissue structures or lesions.

[0073] The maximum grayscale value refers to the maximum value of the grayscale values ​​of all pixels in the image, which represents the brightest part of the image and usually corresponds to the white or nearly white area. The size of the above maximum grayscale value can reflect the bright details and contrast of the image. A larger maximum grayscale value means that there are brighter areas in the image, which helps to display high-density tissue structures or lesions.

[0074] The minimum grayscale value and maximum grayscale value of an image represent the darkest and brightest parts of the image, respectively, and are important for understanding the contrast and brightness distribution of the image, as well as subsequent image processing and analysis. In practical applications, it is necessary to select the appropriate grayscale range and processing method based on specific diagnostic needs and image characteristics.

[0075] Furthermore, in the present embodiment, the analysis and adjustment method of the patient's medical image is only an optional condition of the present embodiment. In one or some other embodiments, the medical image adjustment method can be replaced according to the medical image diagnosis requirements and image standards. Different medical image diagnosis tasks require attention to different image features. By adjusting the analysis and adjustment methods according to specific diagnostic requirements, it can be ensured that the extracted and displayed medical image information meets the basic requirements, thereby improving the accuracy of the diagnosis results.

[0076] S4. A medical image feature fusion mechanism is set up according to the medical image information, and medical image feature information is obtained according to the medical image feature fusion mechanism. The specific steps and implementation contents are as follows:

[0077] An image processing and storage module is configured in the Internet medical image diagnosis management system. The image processing and storage module sets a medical image feature fusion mechanism based on the processed medical image information. In the embodiment, the medical image feature fusion mechanism mainly includes: a neural network branch feature information model and an image feature fusion model, and its specific contents are as follows:

[0078] The image processing and storage module in the embodiment mainly includes an image processing module and an information storage module.

[0079] The feature information of different branches in the artificial neural network is mainly obtained through the image processing module. In an optional embodiment, the feature information set of different branches in the artificial neural network is obtained through the neural network branch feature information model.

[0080] In the image processing module, features are extracted from different branches of the artificial neural network to convert the visual information in the medical image into a quantifiable and analyzable data set. The above process not only pursues the objectivity of the information, but also ensures the accuracy and effectiveness of the conversion from medical images to data.

[0081] In view of the ability to learn the importance of each channel through jump connections in the artificial neural network model and its excitation mechanism, the present invention introduces a neural network branch feature information model. This model can deeply mine and separate the feature information sets of different branches in the artificial neural network. In the embodiment, the branches can be different levels, different convolution kernel sets or specific functional modules in the artificial neural network, which respectively capture the feature information of different levels and dimensions in the medical image.

[0082] The neural network branch feature information model can systematically and selectively extract the feature information contained in different branches, thereby constructing a rich, multi-dimensional feature representation space, which not only reflects the essential attributes of medical images, but also reveals the intrinsic connections and interactions between different features, providing a solid foundation for subsequent tasks such as data analysis, pattern recognition or disease diagnosis.

[0083] The above neural network branch feature information model satisfies the following relationship:

[0084]

[0085] Among them, n It represents the feature information set of different branches in the artificial neural network. L represents the first feature size of the nth branch, Y represents the second feature size of the nth branch, m represents the total number of nodes in the nth branch, and PMI m (l, y) represents the feature information of the mth node in the branch, λ m Represents the weight corresponding to the mth node in the branch.

[0086] The characteristic size of an artificial neural network model refers to the dimensions of the output information of different layers or neurons in the artificial neural network model, including but not limited to height, width and depth. The above dimensions jointly determine the size and amount of information output by different layers or neurons.

[0087] The first and second characteristic sizes of different branches of the artificial neural network refer to the fact that different branches are processed by different convolution kernels, step sizes or padding strategies, so the first characteristic size and the second characteristic size, that is, the output characteristic size of the first one or two layers, make it closest to the input layer or a specific starting point of the network.

[0088] The total number of nodes contained in different branches of an artificial neural network. In order to achieve specific functions or process specific information, different branches in the neural network further contain different numbers of layers and different types of nodes, and different branches have different node configurations.

[0089] In artificial neural networks, the weights corresponding to different nodes in different branches refer to the specific numerical parameters of the connection between each node and the nodes in the previous layer in different branches of the neural network. The above weights play a vital role in the training and learning process of the neural network, determining the strength and direction of signal transmission between nodes. Each node receives input signals from multiple nodes in the previous layer. The input signals are weighted and summed by their respective weights, and then a bias term is added. Finally, the output of the node is obtained through the activation function. Therefore, the size and positive and negative of the weight directly affects the output value of the node, and then affects the performance of the entire network.

[0090] The contribution of each node to the final output information is further calculated and its influence on the feature map fusion process, that is, channel-by-channel multiplication fusion is used to calculate the contribution weights of different nodes in the artificial neural network to ensure the accuracy of the information set of the fused medical image features.

[0091] The weights corresponding to different nodes in different branches above satisfy the following relationship:

[0092]

[0093] Among them, λ m represents the weight corresponding to the mth node in the artificial neural network, and softmax represents the softmax function that makes the sum of the weights equal to 1. Represents the feature information of the mth node in the artificial neural network after multi-dimensional compression. represents the channel-by-channel multiplication symbol, l represents the primary feature size of different branches, y represents the secondary feature size of different branches, and T represents the medical image feature information matrix of different branches.

[0094] In artificial neural networks, feature information after multi-dimensional compression of different nodes refers to the information obtained after compressing the feature representations of different nodes in the network in order to reduce computational complexity, reduce storage requirements or increase processing speed during the neural network processing. This compression usually involves multiple dimensions, including but not limited to the number, accuracy, spatial resolution, etc. Based on this, the number of features can be reduced, the amount of calculation in the neural network can be reduced, and the processing speed can be increased. The compressed feature information takes up less storage space, which is conducive to the deployment of neural network models in resource-constrained environments, thereby improving the generalization ability of artificial neural network models.

[0095] When analyzing medical images based on artificial neural network models, the medical image feature information matrix of different branches refers to the matrix form of feature information extracted and represented for medical image data on different branches of the network. The above feature information matrix is ​​the basis for neural networks to perform image recognition, classification, detection and other tasks. The above matrix includes but is not limited to the texture, shape, edge, grayscale distribution and other features of medical images.

[0096] Next, the medical image feature information is fused through the image processing module. In another optional embodiment, the fusion information of the medical image features is obtained based on the feature information sets of different branches and the image feature fusion model.

[0097] The feature information sets of different branches are sent to the image processing module for fusion. The feature information of different branches is integrated according to the image feature fusion model. The output of the fusion operation function is the final output medical image feature fusion information, that is, the medical image feature information is obtained. The final output medical image feature fusion information integrates the feature information from different branches and provides a comprehensive feature representation for subsequent medical image processing, analysis or diagnosis.

[0098] The above image feature fusion model satisfies the following relationship:

[0099]

[0100] Among them, FFI represents the final output of medical image feature fusion information, fix represents the fusion operation function of medical image features, N represents the number of branches of medical images in the artificial neural network, g represents the activation function of image feature information, and ffi i Represents the feature information set of the i-th branch in the artificial neural network.

[0101] In order to further verify the feasibility of the image feature fusion model in the image processing module, the image feature fusion model in the image processing module is compared and analyzed with the existing data fusion model in the embodiment, and the specific contents are as follows:

[0102] Based on the historical medical imaging information and diagnosis records of the Internet medical imaging diagnosis management system, the accuracy of different feature fusion models is compared. In the embodiment, three feature fusion models are introduced, and the results of the four feature fusion models are compared. For the comparison information of the evaluation index, please see Table 1.

[0103] Table 1 Accuracy comparison data of different feature fusion models

[0104]

[0105]

[0106] Based on the comparison results, it can be seen that the accuracy, precision, recall rate and F1 value of the image feature fusion model in the image processing module are higher than those of the existing feature fusion model, highlighting the unique advantages of the image feature fusion model of the present invention in processing complex medical image data, and providing a new perspective and strong support for optimizing the medical image analysis process and improving the accuracy of disease diagnosis.

[0107] The information storage module in the image processing and storage module in the embodiment stores the output medical image fusion feature information.

[0108] The information storage module saves the fused medical image feature information safely and efficiently into the database. The above process not only ensures the persistence of the fused feature information of medical images, but also provides a solid basic data support for subsequent clinical diagnosis, in-depth analysis and scientific research.

[0109] At the same time, in order to meet the growing demand for data storage, the Internet Medical Imaging Diagnosis Management System has better availability and scalability. The better availability of the system ensures that the stored data can remain continuously accessible even in the face of system failures or maintenance activities, thereby not interrupting the normal operation of medical services. On the other hand, better scalability enables the system to flexibly respond to the surge in data volume, maintain high performance and stability through seamless expansion, and support long-term, large-scale data management and utilization in hospitals or research institutions.

[0110] Through its efficient and reliable storage mechanism, the information storage module not only ensures the secure storage of medical image fusion feature information, but also promotes the in-depth mining and value maximization of medical data, making important contributions to the digital transformation and development of the medical industry.

[0111] Furthermore, in the present embodiment, the method for acquiring medical image feature information is only an optional condition of the present embodiment. In one or some other embodiments, the method for acquiring medical image feature information may be adjusted according to the actual operation of the system and the actual conditions of medical image diagnosis. Different medical imaging devices, different disease types and different diagnostic goals require different feature extraction strategies. Adjusting the acquisition method system can better adapt to changes in actual needs and ensure the effectiveness and pertinence of feature information.

[0112] S5. Provide medical diagnosis services to patients by combining standardized medical images, medical image feature information and medical image information. The specific steps and implementation contents are as follows:

[0113] In the embodiment, the cloud image receiving and forwarding module transmits the standardized medical image to the image processing and storage module, which analyzes the medical diagnostic image and uses computer vision and image processing technology to extract key feature information in the medical image, including but not limited to key information such as shape, size, density, and signal strength. Based on the received standardized medical image and medical image feature information, the image intelligent diagnosis module compares the key feature information with the standard feature information in the system database, and then preliminarily determines the type and nature of the patient's lesion.

[0114] The intelligent imaging diagnosis module is also equipped with advanced medical imaging visualization equipment. During the diagnosis process, the medical imaging visualization equipment provides doctors with an intuitive and clear imaging display platform, and deeply integrates the multi-dimensional clinical information of patients, so that they can seamlessly switch between patients' imaging and medical history records, observe the clinical manifestations of patients, and refer to laboratory test data for comprehensive analysis from all angles. It not only speeds up the imaging diagnosis process, improves the accuracy and efficiency of diagnosis, but also promotes the optimal allocation of medical resources, providing patients with more personalized and accurate treatment plans. It also marks an important step for medical diagnostic technology to move towards intelligence and informatization.

[0115] The image intelligent diagnosis module not only has powerful visual image analysis functions, but also can instantly generate detailed and standardized diagnostic reports and feedback information. This module integrates the precise diagnostic results of medical images, detailed medical records of patients, and corresponding medical image feature information, and integrates the above information into an intuitive and easy-to-understand written report through intelligent processing. The above report not only accurately describes the abnormal structure found in the medical image, but also clearly points out the location of the abnormality and gives professional diagnostic opinions. The report format of the image intelligent diagnosis module is both unified and standardized, strictly following the professional standards and requirements of medical image diagnosis, greatly improving the efficiency and accuracy of doctors in reading and understanding diagnostic results, and providing strong support for clinical decision-making.

[0116] The present invention provides an Internet-based medical imaging diagnosis method, which integrates medical imaging processing and analysis technologies, aims to accurately extract and dynamically load relevant features of medical images, and realize efficient organization and management of medical imaging information through intelligent classification algorithms. It also uses Internet technology to build an online diagnosis and treatment system that is comprehensive, intuitive, objective and accurate. It can not only significantly improve the accuracy and efficiency of medical imaging diagnosis, but also promote the development of medical imaging management technology, and promote the medical diagnosis system to move towards a higher level of intelligence, digitization and networking, providing patients with a more convenient, accurate and personalized medical service experience.

[0117] More importantly, this method fully explores and utilizes the value of standardized medical images, image feature information and precise medical image information, ensuring the scientificity, accuracy and timeliness of the medical diagnosis process, bringing patients a more efficient, personalized and high-quality medical service experience, and helping to promote the intelligent development of the medical industry.

[0118] See also Figure 4In an optional embodiment, in order to efficiently execute the Internet-based medical imaging diagnosis method provided by the present invention, the present invention also provides an Internet-based medical imaging diagnosis system. The above-mentioned Internet-based medical imaging diagnosis system includes a medical imaging acquisition module, a cloud-based imaging receiving and forwarding module, an imaging processing and storage module, and an imaging intelligent diagnosis module. The medical imaging acquisition module, the cloud-based imaging receiving and forwarding module, the imaging processing and storage module, and the imaging intelligent diagnosis module are interconnected. The above-mentioned system executes the specific steps of the relevant embodiments of the Internet-based medical imaging diagnosis method provided by the present invention. The Internet-based medical imaging diagnosis system of the present invention has a complete structure, is objective and stable, and can efficiently execute the Internet-based medical imaging diagnosis method of the present invention, thereby improving the overall applicability and practical application ability of the present invention.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A medical imaging diagnosis and management method based on the Internet. It is characterized in that The steps include: Obtain the patient's initial medical imaging information through the Internet medical imaging diagnosis management system; Performing information adjustment on the initial medical image information to obtain processed medical image information; Building an image correction model based on the medical image information, and using the image correction model to analyze and adjust the patient's medical image to obtain a standardized medical image; Setting a medical image feature fusion mechanism according to the medical image information, and obtaining medical image feature information according to the medical image feature fusion mechanism; The standardized medical images, the medical image feature information and the medical image information are combined to provide medical diagnosis services for patients.

2. The Internet-based medical imaging diagnosis management method according to claim 1, It is characterized in that The step of adjusting the initial medical image information to obtain processed medical image information includes: Configuring a cloud image receiving and forwarding module in the Internet medical image diagnosis management system; An artificial neural network model is set in the cloud image receiving and forwarding module, and the initial medical image information is adjusted using the artificial neural network model to obtain processed medical image information.

3. The Internet-based medical imaging diagnosis and management method according to claim 2, It is characterized in that The medical image information after the processing satisfies the following relationship: Among them, PMI represents the processed medical image information, R represents the artificial neural network activation model, m represents the total number of neurons in the artificial neural network, and I i represents the medical image information input by the i-th neuron, represents the connection weight of the i-th neuron in the artificial neural network, and τ represents the bias value of the artificial neural network.

4. The Internet-based medical imaging diagnosis and management method according to claim 1, It is characterized in that The setting of the medical image feature fusion mechanism according to the medical image information includes: Configuring an image processing and storage module in the Internet medical image diagnosis management system; The image processing and storage module sets a medical image feature fusion mechanism according to the processed medical image information, and the medical image feature fusion mechanism includes a neural network branch feature information model and an image feature fusion model.

5. The Internet-based medical imaging diagnosis and management method according to claim 4, It is characterized in that The neural network branch feature information model satisfies the following relationship: Among them, n It represents the feature information set of different branches in the artificial neural network. L represents the first feature size of the nth branch, Y represents the second feature size of the nth branch, m represents the total number of nodes in the nth branch, and PMI m (l, y) represents the feature information of the mth node in the branch, λ m Represents the weight corresponding to the mth node in the branch.

6. The Internet-based medical imaging diagnosis and management method according to claim 5, It is characterized in that The image feature fusion model satisfies the following relationship: Among them, FFI represents the final output of medical image feature fusion information, fix represents the fusion operation function of medical image features, N represents the number of branches of medical images in the artificial neural network, g represents the activation function of image feature information, and ffi i Represents the feature information set of the i-th branch in the artificial neural network.

7. The Internet-based medical imaging diagnosis and management method according to claim 4, It is characterized in that The obtaining of medical image feature information according to the medical image feature fusion mechanism includes: Obtaining a feature information set of different branches in the artificial neural network through the neural network branch feature information model; Based on the feature information sets of different branches and the image feature fusion model, fusion information of medical image features is obtained.

8. The Internet-based medical imaging diagnosis and management method according to claim 1, It is characterized in that The constructing of an image correction model based on the medical image information, and using the image correction model to analyze and adjust the patient's medical image to obtain a standardized medical image comprises: Configuring an image intelligent diagnosis module in the Internet medical image diagnosis management system; Building an image correction model in the image intelligent diagnosis module based on the medical image information; The image correction model is used to analyze and adjust the patient's medical images to obtain standardized medical images.

9. The Internet-based medical imaging diagnosis and management method according to claim 8, It is characterized in that The image correction model satisfies the following relationship: Among them, J C represents the gray value of the image after standardization, σ represents the comprehensive influence coefficient of abnormal noise in the image, J represents the initial gray value of the image, and J min Indicates the minimum gray value of the image, J max Indicates the maximum grayscale value of the image.

10. A medical imaging diagnosis management system based on the Internet. It is characterized in that The system includes a medical image acquisition module, a cloud image receiving and forwarding module, an image processing and storage module, and an image intelligent diagnosis module. The medical image acquisition module, the cloud image receiving and forwarding module, the image processing and storage module, and the image intelligent diagnosis module are interconnected to execute the Internet-based medical image diagnosis management method as described in any one of claims 1 to 9.

Images