Emphysema grade detection method, device, equipment and storage medium

Abstract

The present invention discloses a method, device, equipment and storage medium for detecting emphysema grade, the method comprising: obtaining image features of a lung image; performing convolution processing on the image features of the lung image to obtain an attention feature matrix; and predicting the emphysema grade of the lung image using the attention feature matrix. The present invention determines the emphysema grade by using the attention feature matrix after processing the lung image, which can greatly reduce the time cost of pulmonary function examinations and reduce the risk of missed diagnosis, thereby improving the detection efficiency and accuracy of the emphysema grade.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method, device, equipment and storage medium for detecting emphysema grade. Background Art

[0002] Chronic obstructive pulmonary disease (emphysema) is a disease that seriously threatens human health. At present, pulmonary function tests are a method to determine emphysema and the grade of emphysema. The clinical significance of maximum minute ventilation (MVV): This test is essentially a test of ventilation reserve capacity. It is used to measure the elasticity of thoracic lung tissue, airway resistance, and respiratory muscle strength. In medicine, the ratio of the measured value to the theoretical predicted value is often used to express its size. Normal is greater than 80%, and less than 60% is abnormal-reduced ventilation reserve capacity. Residual volume (RV) / total lung capacity (TCL): The ratio of residual volume to total lung capacity is often used to determine whether there is emphysema. <25% is normal; 26% to 35% is mild emphysema; 36% to 45% is moderate emphysema; 46% to 55% is severe emphysema; >55% is extremely severe emphysema. The RV / TCL ratio is related to age. The residual volume increases with age, but generally should not exceed 35%.

[0003] The method of grading emphysema using lung function tests can be found in the Global Initiative for Chronic Obstructive Pulmonary (Emphysema) Disease (GOLD). The emphysema grades include: GOLD grade 0 (absent), GOLD grade 1 (mild), GOLD grade 2 (moderate), GOLD grade 3 (severe) and GOLD grade 4 (extremely severe).

[0004] However, the procedures for pulmonary function tests are cumbersome, the time cost of the tests is high, and there is often a risk of missed diagnosis, resulting in low efficiency and accuracy in detecting the grade of emphysema. Summary of the invention

[0005] The main purpose of the present invention is to provide a method, device, equipment and storage medium for detecting emphysema grade, aiming to solve the problem of low efficiency and low accuracy of existing emphysema grade detection.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a method for detecting emphysema grade, comprising:

[0007] Obtaining image features of a lung image;

[0008] Performing convolution processing on the image features of the lung image to obtain an attention feature matrix;

[0009] The emphysema grade of the lung image is predicted using the attention feature matrix.

[0010] Optionally, the using the attention feature matrix to predict the emphysema grade of the lung image includes:

[0011] Obtaining a pulmonary emphysema region of the lung image using the attention feature matrix and the region candidate network;

[0012] The classification network is used to predict the grade of the emphysema area to obtain the emphysema grade of the lung image.

[0013] Optionally, performing convolution processing on the image features of the lung image to obtain an attention feature matrix includes:

[0014] Performing a first convolution process on the image features of the lung image to obtain a plurality of benchmark attention features, wherein the elements in the benchmark attention features are benchmark attentions corresponding to each feature point in the image features;

[0015] Performing a second convolution process on the image features to obtain an attention coefficient matrix, wherein the elements in the attention coefficient matrix are benchmark attention coefficients corresponding to each of the feature points in the image features, and the number of channels of the attention coefficient matrix is ​​the same as the number of the benchmark attention features;

[0016] Based on the image features, each of the benchmark attention features and the attention coefficient matrix, the attention features of each of the feature points are obtained, and the attention feature matrix is ​​composed of the attention features of each of the feature points.

[0017] Optionally, obtaining the attention feature of each of the feature points based on the image feature, each of the benchmark attention features and the attention coefficient matrix includes:

[0018] Based on the image feature and each of the benchmark attention features, obtaining a benchmark feature vector corresponding to each benchmark attention feature, wherein the benchmark feature vector is obtained by summing the products of the image feature and corresponding elements in the benchmark attention feature;

[0019] Multiplying each of the benchmark feature vectors by the attention coefficient matrix to obtain a product matrix for each benchmark feature vector;

[0020] The product values ​​of the same feature points in the product matrix of each benchmark feature vector are added together to obtain the attention features of each feature point.

[0021] Furthermore, the method further comprises:

[0022] Determining the emphysema grade of the lung function data according to the lung function data of the lung corresponding to the lung image;

[0023] Determine a first volume of the emphysema region and a second volume of the small airway lesion region according to the lung image respectively;

[0024] Correcting the emphysema level of the lung function data according to the emphysema level of the lung image;

[0025] classifying the corrected emphysema grade according to the first volume and the second volume to obtain a first emphysema grade category;

[0026] A case report is generated based on the lung function data, the corrected emphysema grade, and the first emphysema grade category.

[0027] Furthermore, after determining the first volume of the emphysema region and the second volume of the small airway lesion region according to the lung image, the method further includes:

[0028] Determining whether the emphysema level of the lung image is the same as the emphysema level of the lung function data;

[0029] If they are the same level, classifying the emphysema level of the pulmonary function data or the emphysema level of the lung image according to the first volume and the second volume to obtain a second emphysema level category;

[0030] generating a case report according to the lung function data, the emphysema grade of the lung function data or the emphysema grade of the lung image, and the second emphysema grade category;

[0031] The step of correcting the emphysema level of the lung function data according to the emphysema level of the lung image;

[0032] If they are different levels, the emphysema level of the lung function data is corrected according to the emphysema level of the lung image.

[0033] Optionally, before determining the first volume of the emphysema region and the second volume of the small airway lesion region according to the lung image, the method further includes:

[0034] Determining the number of slices, scanning slice thickness and slice spacing of the lung image;

[0035] Determining a first volume of an emphysema region and a second volume of a small airway lesion region according to the lung image comprises:

[0036] Determine the first volume according to the emphysema area and the number of slices, the scanning slice thickness and the slice spacing;

[0037] The second volume is determined according to the small airway lesion area, the number of layers, the scanning layer thickness, and the layer spacing.

[0038] Optionally, classifying the corrected emphysema grade according to the first volume and the second volume to obtain a first emphysema grade category includes:

[0039] determining a ratio of the first volume to the second volume;

[0040] The corrected lung function grade is classified according to the ratio to obtain a first emphysema grade category.

[0041] In a second aspect, the present invention provides a device for detecting emphysema grade, the device comprising:

[0042] An acquisition module, used for acquiring image features of a lung image;

[0043] A processing module, configured to perform convolution processing on the image features of the lung image to obtain an attention feature matrix;

[0044] A prediction module is used to predict the emphysema grade of the lung image by using the attention feature matrix.

[0045] Optionally, the processing module includes:

[0046] A processing unit, configured to perform a first convolution process on the image features of the lung image to obtain a plurality of reference attention features, wherein the elements in the reference attention features are reference attention points corresponding to the feature points in the image features;

[0047] The processing unit is used to perform a second convolution process on the image feature to obtain an attention coefficient matrix, wherein the elements in the attention coefficient matrix are benchmark attention coefficients corresponding to each of the feature points in the image feature, and the number of channels of the attention coefficient matrix is ​​the same as the number of the benchmark attention features;

[0048] An acquisition unit is used to obtain the attention features of each of the feature points based on the image features, each of the baseline attention features and the attention coefficient matrix, wherein the attention feature matrix is ​​composed of the attention features of each of the feature points.

[0049] Optionally, the acquisition unit is used to obtain a benchmark feature vector corresponding to each benchmark attention feature based on the image feature and each benchmark attention feature, the benchmark feature vector being obtained by the sum of the products of corresponding elements in the image feature and the benchmark attention feature; multiplying each benchmark feature vector with the attention coefficient matrix to obtain a product matrix for each benchmark feature vector; and adding the product values ​​of the same feature point in the product matrix for each benchmark feature vector to obtain the attention feature of each feature point.

[0050] Optionally, the prediction module is used to obtain the emphysema region of the lung image using the attention feature matrix and the region candidate network; and to perform grade prediction on the emphysema region using a classification network to obtain the emphysema grade of the lung image.

[0051] Furthermore, the device also includes:

[0052] a determination module, configured to determine the emphysema grade of the lung function data according to the lung function data of the lung corresponding to the lung image;

[0053] The determination module is further used to determine a first volume of an emphysema region and a second volume of a small airway lesion region according to the lung image;

[0054] a correction module, configured to correct the emphysema level of the lung function data according to the emphysema level of the lung image;

[0055] a classification module, configured to classify the corrected emphysema grade according to the first volume and the second volume to obtain a first emphysema grade category;

[0056] A generating module is used to generate a case report based on the lung function data, the corrected emphysema grade and the first emphysema grade category.

[0057] Furthermore, the device also includes:

[0058] A judgment module, used for judging whether the emphysema level of the lung image is the same as the emphysema level of the lung function data;

[0059] The classification module is further configured to classify the emphysema grade of the pulmonary function data or the emphysema grade of the lung image according to the first volume and the second volume, if they are the same grade, to obtain a second emphysema grade category;

[0060] The generating module is further configured to generate a case report according to the pulmonary function data, the emphysema grade of the pulmonary function data or the emphysema grade of the lung image, and the second emphysema grade category;

[0061] The correction module is specifically used to correct the emphysema level of the lung function data according to the emphysema level of the lung image if the levels are different.

[0062] In a third aspect, the present invention further proposes an emphysema level detection device, which comprises: a memory, a processor, and an emphysema level detection program stored in the memory and executable on the processor, wherein the emphysema level detection program is executed by the processor to perform the steps of the emphysema level detection method.

[0063] In a fourth aspect, the present invention further proposes a computer-readable storage medium, on which a pulmonary emphysema grade detection program is stored, and when the pulmonary emphysema grade detection program is executed by a processor, the steps of the pulmonary emphysema grade detection method are implemented.

[0064] The present invention provides a method, device, equipment and storage medium for detecting emphysema level. After acquiring the image features of a lung image, convolution processing can be performed on the image features of the lung image to obtain an attention feature matrix. The attention feature matrix is ​​used to predict the emphysema level of the lung image. The emphysema level can be determined through the lung image, which can greatly reduce the time cost of lung function examinations and reduce the risk of missed diagnosis, thereby improving the detection efficiency and accuracy of the emphysema level. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a schematic diagram of the structure of the hardware operating environment involved in the embodiment of the present invention;

[0066] Figure 2 A schematic diagram of a flow chart of a method for detecting emphysema grade according to an embodiment of the present invention;

[0067] Figure 3 is a flow chart of step S20 in a method for detecting emphysema grade according to an embodiment of the present invention;

[0068] Figure 4 This is a flow chart of step S23 in a method for detecting emphysema grade according to an embodiment of the present invention;

[0069] Figure 5 Schematic diagram of functional modules of an emphysema grade detection device according to an embodiment of the present invention.

[0070] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0071] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0072] like Figure 1 As shown, Figure 1 It is a schematic diagram of the device structure of the hardware operating environment involved in the embodiment of the present invention.

[0073] It should be noted that the emphysema level detection method provided in an embodiment of the present invention can be used to improve the detection efficiency and accuracy of emphysema level. The executor of the method can be any emphysema level detection device. For example, the emphysema level detection method can be executed by a terminal device or a server or other processing device, wherein the terminal device can be a user equipment (User Equipment, UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (Personal Digital Assistant, PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc., and no specific restrictions are made here.

[0074] like Figure 1 As shown, the emphysema grade detection device may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0075] Those skilled in the art will understand that Figure 1 The device structure shown in the figure does not constitute a limitation on the emphysema grade detection device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0076] like Figure 1As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and an emphysema grade detection program. The operating system is a program for managing and controlling the hardware and software resources of the device, and supports the operation of the emphysema grade detection program and other software or programs.

[0077] exist Figure 1 In the device shown, the user interface 1003 is mainly used for data communication with the client; the network interface 1004 is mainly used for establishing a communication connection with the server; and the processor 1001 can be used to call the emphysema grade detection program stored in the memory 1005 and execute the emphysema grade detection method of the embodiment of the present invention.

[0078] Furthermore, a method for detecting the grade of emphysema according to an embodiment of the present invention is provided, such as Figure 2 As shown, the method for detecting the grade of emphysema includes:

[0079] S10: acquiring image features of the lung image;

[0080] S20: performing convolution processing on the image features of the lung image to obtain an attention feature matrix to obtain the image features of the lung image;

[0081] S30: Predicting the emphysema grade of the lung image using the attention feature matrix.

[0082] In some possible implementations, the lung image of the embodiment of the present invention can be directly obtained by taking CT (Computed Tomography) or can be transmitted through other electronic devices.

[0083] When a lung image is obtained, the image features of the lung image are first obtained, and the image features can be obtained by performing feature extraction processing on the lung image. The convolution operation of the lung image can be performed through a residual network, a pyramid feature extraction network, or other convolutional neural networks to obtain the image features of the lung image. The image features are represented in the form of a matrix or a vector.

[0084] When the image features of the lung image are obtained, the image features can be used to obtain an attention feature matrix of the lung image.

[0085] Figure 3 A flowchart of step S20 of a method for detecting emphysema grade according to an embodiment of the present invention is shown. In which, the image features of the lung image are subjected to convolution processing to obtain an attention feature matrix, including:

[0086] S21: performing a first convolution process on the image features of the lung image to obtain a plurality of benchmark attention features, wherein the elements in the benchmark attention features are benchmark attentions corresponding to each feature point in the image features;

[0087] S22: performing a second convolution process on the image feature to obtain an attention coefficient matrix, wherein the elements in the attention coefficient matrix are benchmark attention coefficients corresponding to each of the feature points in the image feature, and the number of channels of the attention coefficient matrix is ​​the same as the number of the benchmark attention features;

[0088] S23: Based on the image features, each of the benchmark attention features and the attention coefficient matrix, the attention features of each of the feature points are obtained, and the attention feature matrix is ​​composed of the attention features of each of the feature points.

[0089] In the embodiment of the present invention, steps S21 and S22 may be performed simultaneously or separately, and the order of the two steps is not specifically limited.

[0090] Through step S21, the first convolution processing can be performed on the image features to obtain multiple benchmark attention features, wherein the first convolution kernel can be used to perform convolution processing on the image features in the feature direction to obtain a preset number of benchmark attention features. The preset number can be a preset value, such as 16, 32, etc., which can be set by the technicians in this field according to the needs. The elements in the obtained benchmark attention features are the benchmark attentions corresponding to each feature point in the image features. That is, the benchmark attention for each feature point can be obtained through the first convolution processing, and the benchmark attention can be in the form of a feature vector. In addition, the length of each benchmark attention feature obtained is the same as the length of the image feature, and the width of the benchmark attention feature is the same as the width of the image feature.

[0091] Among them, the feature points described in the embodiment of the present invention can be pixel points of the lung lobe segmentation image, and the first convolution kernel can be a 1*1 convolution kernel, but it is not a specific limitation of the present invention.

[0092] In step S21, a plurality of benchmark attention features can be obtained through the first convolution process, each benchmark attention feature can include benchmark attention corresponding to each feature point in the image feature, and each benchmark attention feature obtained is different. Moreover, the length of the benchmark attention feature is the same as the length of the corresponding feature point in the image feature, and the width in the benchmark attention feature is the same as the width of the corresponding feature point in the image feature, that is, the benchmark attention corresponding to the feature point of the benchmark attention feature corresponds one-to-one to the feature value of each feature point in the image feature. For example, for an image feature with a dimension of a×b×c, the dimension of each benchmark attention feature is a×b×1.

[0093] In addition, through step S22, a second convolution process can be performed on the image features to obtain an attention coefficient matrix, wherein the second convolution kernel can be used to perform convolution process on the image features to obtain an attention coefficient matrix, and the elements in the attention coefficient matrix are the benchmark attention coefficients corresponding to the feature points of the image features. Moreover, the number of channels of the attention coefficients corresponding to each feature point in the benchmark attention coefficient matrix is ​​the same as the number of benchmark attention features. That is, each element in the benchmark attention coefficient matrix can represent multiple benchmark attention coefficients corresponding to the corresponding feature points, and the number of the benchmark attention coefficients is the same as the number of benchmark attention features. That is, each benchmark attention for a feature point can correspond to a corresponding benchmark attention coefficient.

[0094] The second convolution kernel may be a 1*1 convolution kernel, which is not specifically limited in the embodiment of the present invention. The first convolution kernel and the second convolution kernel in the embodiment of the present invention may be different convolution kernels.

[0095] In step S22, in the attention coefficient matrix obtained by performing the second convolution process on the image feature, the elements in each attention coefficient matrix are multiple attention coefficients for different feature points. That is, through the second convolution process, an attention coefficient matrix representing the attention coefficients of each feature point can be obtained, and each attention coefficient matrix can include attention coefficients corresponding to each feature point in the image feature. And the length of the obtained attention coefficient matrix is ​​the same as the length of the image feature, the width of the attention coefficient matrix is ​​the same as the width of the image feature, and the obtained attention coefficient matrix can be expressed as a×b×n, where n is the number of reference attention features.

[0096] After obtaining the attention coefficient matrix and the baseline attention feature, step S23 can be performed based on the two types of feature matrices obtained.

[0097] Among them, in an embodiment of the present invention, the obtaining of a benchmark feature vector corresponding to each benchmark attention feature based on the image feature and each benchmark attention feature includes: multiplying the feature vector corresponding to each feature point in the image feature with the first attention of the same feature point in the benchmark attention feature to obtain a first product for each feature point in the benchmark attention feature; adding the first products for each feature point in the benchmark attention feature to obtain a benchmark feature vector corresponding to the benchmark attention feature.

[0098] Among them, after obtaining the benchmark attention feature, the benchmark attention feature can be multiplied by the vector value corresponding to the corresponding feature point in the image feature to obtain the first product of the feature point for the benchmark attention feature. For example, image feature A can be expressed as [a1, b1, c1], a benchmark attention feature B can be expressed as [a2, b2, c2], and another benchmark attention feature C can be expressed as [a3, b3, c3]. Then step S23 can obtain the first product of the three feature points for the benchmark attention feature B, namely a1a2, b1b2 and c1c2, and the first product of the three feature points for the benchmark attention feature C, namely a1a3, b1b3 and c1c3. Similarly, the first product value of each feature point for each benchmark attention feature can be obtained.

[0099] Furthermore, the first product for each feature point in the baseline attention feature can be summed to obtain a baseline feature vector corresponding to the baseline attention feature. That is, after obtaining the first product value of each identical feature point in a baseline attention feature, the first product of each feature point can be summed to obtain a baseline feature vector. For example, for the baseline attention feature B, the sum vector a1a2, b1b2 and c1c2 can be obtained, that is, the baseline feature vector a1a2+b1b2+c1c2 for the baseline attention feature B can be obtained. Similarly, the baseline feature vector a1a3+b1b3+c1c3 for the baseline attention feature C can be obtained. By analogy, the baseline feature vector of each baseline attention feature can be obtained.

[0100] In the above configuration of the embodiment of the present invention, it is not necessary to calculate the correlation between each feature point and other feature points and construct a correlation tensor about each feature point, which greatly improves the calculation speed and reduces the consumption of calculation resources.

[0101] In addition, after obtaining the attention coefficient matrix, the attention features of each feature point can be obtained through step S23. Figure 4 A flowchart of step S23 in a lesion area detection method according to an embodiment of the present invention is shown, wherein the step of obtaining the attention feature of each feature point based on the image feature, each of the benchmark attention features and the attention coefficient matrix may include:

[0102] S231: Based on the image feature and each of the benchmark attention features, obtaining a benchmark feature vector corresponding to each benchmark attention feature, wherein the benchmark feature vector is obtained by summing the products of the image feature and corresponding elements in the benchmark attention feature;

[0103] S232: multiplying each of the reference feature vectors by the attention coefficient matrix to obtain a product matrix for each reference feature vector;

[0104] S233: Add the product values ​​of the same feature point in the product matrix of each attention coefficient matrix to obtain the attention feature of the feature point.

[0105] After obtaining the baseline feature vector corresponding to each baseline attention feature, the baseline feature vector can be multiplied with the attention coefficient matrix obtained in step S22 to obtain a product matrix corresponding to the baseline feature vector. Among them, the baseline attention and attention coefficient corresponding to the same feature point can be multiplied. Since each element in the attention coefficient matrix can include multiple attention coefficients for a feature point, and the number of attention coefficients for each feature point is the same as the number of baseline attention features, the product of the one-to-one corresponding attention coefficient and the baseline attention feature can be obtained. Therefore, after the product, multiple product values ​​of the attention coefficient of each feature point and the baseline attention can be obtained. Therefore, the product matrix corresponding to each baseline attention vector can be obtained through the above, and the product matrix includes the above-mentioned multiple product values ​​of each feature point.

[0106] After obtaining the product matrix of each benchmark attention vector, the product values ​​of the corresponding feature points in each product matrix can be added to obtain the attention features for the feature points. As described above, each product matrix can include the product value of the attention coefficient corresponding to each feature point and the benchmark attention. Therefore, the product values ​​of the corresponding attention coefficients of the same point in each product matrix can be added to obtain the attention features for the feature point, thereby obtaining the attention features of each feature point. The attention features obtained from each feature point can constitute an attention feature matrix. According to the above embodiment, the attention of each feature point in the lung lobe segmentation image can be obtained. In the above processing of the embodiment of the present invention, the attention of each feature point is obtained by multiplying the obtained attention coefficient matrix with the benchmark attention vector, and performing a simple addition operation. This process reduces the amount of calculation compared to the calculation process in the prior art.

[0107] Step S20 of the embodiment of the present invention can be implemented by a neural network with an attention mechanism, and the neural network can be a network structure that has been trained to meet the accuracy requirements, and can include a convolution module that performs the first convolution kernel and the second convolution processing, and an operation module that performs the multiplication and addition processing. When achieving global relevance (i.e., determining the attention of each feature point), resource consumption can be reduced, and the task of learning global relevance using the attention module can be completed with significantly fewer resources; and global relevance can be learned on significantly larger inputs.

[0108] When the attention features of each feature point of the lobe segmentation image are obtained, the attention features can be used to perform emphysema area detection. Among them, the embodiment of the present invention can directly obtain the emphysema area through feature processing of the attention feature matrix formed by each feature point, or the attention matrix can be optimized before executing step S30. For example, the optimization process may include inputting the image features and the attention feature matrix composed of the attention features corresponding to each feature point into the residual module for summing processing to obtain an optimized attention feature matrix, and the elements in the optimized attention feature matrix are the optimized attention for each feature point. That is, the attention feature matrix and the image features corresponding to the original input data can be summed to obtain an optimized attention matrix.

[0109] The feature information in the attention feature matrix obtained in the embodiment of the present invention highlights the characteristics of the emphysema area, so when performing emphysema grade prediction through the attention feature matrix, the accuracy of the prediction can be improved. Similarly, performing grade prediction through an optimized attention matrix can further integrate the characteristics of the lung image, thereby further improving the accuracy of detection and grade prediction of emphysema areas. As described in the above embodiments, the present invention realizes that the emphysema grade of the lung image can be predicted using the attention feature matrix or the optimized attention feature matrix. The following description takes the attention feature matrix as an example, and the process of performing grade prediction through the optimized attention matrix will not be repeated.

[0110] The embodiment of the present invention can directly input the attention feature matrix into the region candidate network, through which the emphysema region of the lung image can be detected, and the classification network is used to predict the grade of the detected emphysema region, and the classification network is used to predict the emphysema grade corresponding to the emphysema region in the lung image. The region candidate network and the classification network, as well as the network modules of the first convolution processing and the second convolution processing mentioned above, and the operators that perform multiplication and addition operations constitute a neural network for performing emphysema grade detection in the embodiment of the present invention. The neural network can be a network that has been pre-trained to meet the grade detection accuracy, and the accuracy is higher than 95%. The classification network can also include multiple convolution layers, which is not specifically limited by the present invention.

[0111] Based on the above configuration, the embodiment of the present invention can detect the emphysema area between lung images and automatically detect the level without the need for lung function data detection, which is more convenient and reduces time cost.

[0112] In addition, in an embodiment of the present invention, a case report can also be generated based on the detected emphysema grade. Specifically, it includes: determining the emphysema grade of the pulmonary function data according to the pulmonary function data of the lung corresponding to the lung image; determining the first volume of the emphysema area and the second volume of the small airway lesion area according to the lung image; classifying the emphysema grade of the pulmonary function data according to the first volume and the second volume to obtain the emphysema grade category; generating a case report according to the pulmonary function data, the corrected emphysema grade and the emphysema grade category. To solve the problem that pulmonary function examination cannot give a comprehensive, reasonable and accurate case report. It should be noted here that emphysema is a development result of small airway lesions. Therefore, in the lung image with emphysema, small airway lesions are also included. The embodiment of the present invention can also use the attention features of the above-mentioned lung image to obtain the small airway lesion area, for example, the small airway lesion area is selected through the regional candidate network. The regional candidate network is a trained network that can detect emphysema areas and small airway lesion areas.

[0113] In an embodiment of the present invention, the method for generating a case report may specifically include: determining the emphysema grade of the pulmonary function data according to the pulmonary function data of the lungs corresponding to the lung image; determining a first volume of the emphysema area and a second volume of the small airway lesion area according to the lung image; after determining the first volume of the emphysema area and the second volume of the small airway lesion area according to the lung image, judging whether the emphysema grade of the lung image is the same as the emphysema grade of the pulmonary function data; if they are different levels, correcting the emphysema grade of the pulmonary function data according to the emphysema grade of the lung image; classifying the corrected emphysema grade according to the first volume and the second volume to obtain a first emphysema grade category; generating a case report according to the pulmonary function data, the corrected emphysema grade and the first emphysema grade category. Or if they are the same level, the emphysema level of the pulmonary function data or the emphysema level of the lung image is classified according to the first volume and the second volume to obtain a second emphysema level category; and a case report is generated according to the pulmonary function data, the emphysema level of the pulmonary function data or the emphysema level of the lung image, and the second emphysema level category.

[0114] In an embodiment of the present invention, the method for determining the emphysema grade of the lung image may also be carried out in the following manner, including: determining a first volume of the emphysema region of the lung image and identifying the total volume of the lung image, and determining the emphysema grade of the lung image based on a ratio of the first volume to the total volume.

[0115] For example, four thresholds are set and stored, namely, a first threshold, a second threshold, a third threshold and a fourth threshold. If the ratio of the first volume to the total volume is less than the first threshold, the level of emphysema is GOLD level 0 (non-existent); if the ratio of the first volume to the total volume is between the first threshold and the second threshold, the level of emphysema is GOLD level 1 (mild); if the ratio of the first volume to the total volume is between the second threshold and the third threshold, the level of emphysema is GOLD level 2 (moderate); if the ratio of the first volume to the total volume is between the third threshold and the fourth threshold, the level of emphysema is GOLD level 3 (severe); if the ratio of the first volume to the total volume is greater than the fourth threshold, the level of emphysema is GOLD level 4 (extremely severe). The first threshold, the second threshold, the third threshold and the fourth threshold may be 20%, 30%, 50% and 80% respectively.

[0116] It should be noted that the emphysema grade obtained from the lung function data can only provide an evaluation of the entire lung. The emphysema grade obtained from the lung image can at least provide the emphysema grade of the left lung and the right lung respectively. For example, the emphysema grade of each lung lobe can be obtained by segmenting the lung lobes.

[0117] In an embodiment of the present invention, the lung image includes: at least one of a full inspiratory phase lung image and a full expiratory phase lung image; before determining the first volume of the emphysema region and the second volume of the small airway lesion region according to the lung image, the method further includes: determining the number of layers, scanning layer thickness and inter-layer spacing of the lung image. The number of layers, scanning layer thickness and inter-layer spacing of the full inspiratory phase lung image and the full expiratory phase lung image are the same.

[0118] The method for determining the first volume of the emphysema area and the second volume of the small airway lesion area according to the lung image, respectively, comprises: determining the first volume according to the emphysema area and the number of layers, the scanning layer thickness and the inter-layer spacing; determining the second volume according to the small airway lesion area and the number of layers, the scanning layer thickness and the inter-layer spacing. In an embodiment of the present invention, the full inspiratory phase lung image and the full expiratory phase lung image are lung images of a patient, and the full inspiratory phase lung image is a lung image taken by an inspiratory device when the lung air volume is maximized during deep inhalation. Similarly, the full expiratory phase lung image is a lung image taken by an inspiratory device when the lung air volume is minimized during deep exhalation. The full inspiratory phase lung image and the full expiratory phase lung image can be achieved by radiologists in hospitals with the help of imaging equipment (such as CT).

[0119] Specifically, in an embodiment of the present invention, determining the first volume of the emphysema area according to the lung image includes: using the full inspiratory phase lung image or the full expiratory phase lung image to identify the emphysema area of ​​each layer of the image; determining the first volume according to the emphysema area and the number of layers, scanning layer thickness and layer spacing.

[0120] Specifically, in an embodiment of the present invention, the method for determining the second volume of the small airway lesion area based on the lung image includes: using the full inspiratory phase lung image and the full expiratory phase lung image to identify the small airway lesion area of ​​each layer of the image; determining the second volume based on the small airway lesion area and the number of layers, scanning layer thickness and layer spacing.

[0121] In an embodiment of the present invention, the method for determining the first volume based on the emphysema area, the number of layers, the scanning layer thickness and the inter-layer spacing includes: gridding the emphysema area to obtain a first gridded area of ​​the emphysema area; and obtaining the first volume of the emphysema area based on the first gridded area, the number of layers, the scanning layer thickness and the inter-layer spacing.

[0122] In an embodiment of the present invention, a method for determining the second volume based on the small airway lesion area and the number of layers, the scanning layer thickness and the inter-layer spacing includes: gridding the small airway lesion area to obtain a second gridded area of ​​the small airway lesion area; and obtaining the second volume of the small airway lesion area based on the second gridded area, the number of layers, the scanning layer thickness and the inter-layer spacing.

[0123] In an embodiment of the present invention, the number of layers, scanning layer thickness and inter-layer spacing of the full inspiratory phase lung image or the full expiratory phase lung image are all information inherent in the full inspiratory phase lung image or the full expiratory phase lung image, such as the full inspiratory phase lung image or the full expiratory phase lung image can be a DICOM image. DICOM (Digital Imaging and Communications in Medicine) is digital imaging and communications in medicine, and is an international standard for medical images and related information (ISO 12052). It defines a medical image format that can be used for data exchange and has a quality that meets clinical needs. The number of layers, scanning layer thickness and inter-layer spacing of the image are obtained by reading the DICOM information of the full inspiratory phase lung image or the full expiratory phase lung image.

[0124] In an embodiment of the present invention, the specific method for obtaining the first volume of the emphysema area and the second volume of the small airway lesion area according to the emphysema area and the small airway lesion area, respectively, is: determining the number of image layers, the scanning layer thickness and the inter-layer spacing of the full inspiratory phase lung image or the full expiratory phase lung image; gridding the emphysema area of ​​each layer and the small airway lesion area of ​​each layer to obtain the first grid area of ​​the emphysema area and the second grid area of ​​the small airway lesion area of ​​each layer, and determining the first volume of the emphysema area and the second volume of the small airway lesion area according to the first grid area of ​​the emphysema area and the second grid area of ​​the small airway lesion area of ​​each layer, the number of layers, the scanning layer thickness and the inter-layer spacing.

[0125] Specifically, the first volume of the emphysema region of the left lung is determined according to the first grid area of ​​the emphysema region of each layer, the number of layers, the scanning layer thickness and the inter-layer spacing; the second volume of the small airway lesion region is determined according to the second grid area of ​​the small airway lesion region of each layer, the number of layers, the scanning layer thickness and the inter-layer spacing.

[0126] In an embodiment of the present invention, the method of gridding the emphysema region of each layer and the small airway lesion region of each layer to obtain a first grid area of ​​the emphysema region of each layer and a second grid area of ​​the small airway lesion region, comprises: respectively determining the edges of the emphysema region and the small airway lesion region; respectively drawing grids in the emphysema region and the small airway lesion region, and when the grid contacts the edge, refining the grid, stopping the refinement after the number of grids reaches a set number, and respectively calculating the sum of all grid areas of the emphysema region and the small airway lesion region to obtain the first grid area of ​​the emphysema region and the second grid area of ​​the small airway lesion region.

[0127] Specifically, the gridding process for the emphysema region of each layer to obtain the first grid area of ​​the emphysema region of each layer includes: determining the edge of the emphysema region; drawing a grid in the emphysema region, and when the grid touches the edge, refining the grid, and stopping the refinement after the number of grids reaches a set number, and calculating the sum of all grid areas of the emphysema region to obtain the first grid area of ​​the emphysema region.

[0128] Specifically, the method of gridding the small airway lesion area of ​​each layer to obtain the second grid area of ​​the small airway lesion area of ​​each layer includes: determining the edge of the small airway lesion area; drawing a grid in the small airway lesion area, and when the grid touches the edge, refining the grid, and stopping the refinement after the number of grids reaches a set number, and calculating the sum of all grid areas of the small airway lesion area to obtain the second grid area of ​​the small airway lesion area.

[0129] Since a complete lung includes a left lung and a right lung, in order to enable persons skilled in the art to better understand the present invention, images of the left lung and the right lung are described below respectively.

[0130] The specific method of performing gridding processing on the emphysema region of each layer of the left lung and the small airway lesion region of each layer of the left lung to obtain the first grid area of ​​the emphysema region of each layer of the left lung and the second grid area of ​​the small airway lesion region is as follows: Step 1: respectively determine the edge of the emphysema region of each layer of the left lung and the edge of the small airway lesion region of each layer of the left lung; respectively determine a first standard shape in the emphysema region of each layer of the left lung and the small airway lesion region of each layer of the left lung; Step 2: the first standard shape extends toward the edge of the emphysema region of each layer of the left lung or the edge of the small airway lesion region of each layer of the left lung; Step 3: when several set points of the standard shape touch the edge of the emphysema region of each layer of the left lung or the edge of the small airway lesion region of each layer of the left lung, the first standard shape stops extending and generates a second standard shape outside the standard shape; generate in accordance with steps 2 and 3 in sequence until the number of standard shapes reaches the set number, calculate the areas of all standard shapes, and obtain the first grid area and the first grid area of ​​each layer of the left lung; wherein, The areas of the standard shape, the second standard shape, and the standard shape corresponding to the set quantity decrease in sequence.

[0131] In an embodiment of the present invention, the specific method for gridding the emphysema region of each right lung layer and the small airway lesion region of each right lung layer to obtain the first grid area of ​​the emphysema region of each right lung layer and the second grid area of ​​the small airway lesion region is as follows: Step 1: respectively determine the edge of the emphysema region of each right lung layer and the edge of the small airway lesion region of each right lung layer; respectively determine a first standard shape in the emphysema region of each right lung layer and the small airway lesion region of each right lung layer; Step 2: the first standard shape extends toward the edge of the emphysema region of each right lung layer or the edge of the small airway lesion region of each right lung layer; Step 3: when several set points of the standard shape touch the edge of the emphysema region of each right lung layer or the edge of the small airway lesion region of each right lung layer, the first standard shape stops extending and generates a second standard shape outside the standard shape; generate in accordance with steps 2 and 3 in sequence until the number of standard shapes reaches the set number, Calculate the areas of all standard shapes to obtain the first grid area and the first grid area of ​​each layer of the right lung; wherein the standard shape, the second standard shape, and the standard shape area corresponding to the set number are reduced in sequence. For the value setting of the set number, those skilled in the art can set it according to the accuracy requirements.

[0132] The first gridded area of ​​the emphysema region of each layer is the sum of the first gridded area of ​​the left lung of each layer and the first gridded area of ​​the right lung of each layer. The second gridded area of ​​the small airway lesion region of each layer is the sum of the second gridded area of ​​the left lung of each layer and the second gridded area of ​​the right lung of each layer.

[0133] In an embodiment of the present invention, the first standard shape is at the geometric center of the emphysema region or the small airway lesion region. The standard shape is any one or more of a circle, an ellipse, a rectangle, and a square.

[0134] For example, if the specification shape is a circle, a number of set points can be evenly set on the circumference of the circle, such as 20 set points. When the set points of the specification shape touch the edge of the small airway lesion area, the first specification shape stops extending and generates a second specification shape outside the specification shape; the second specification shape can also be selected from any one of a circle, an ellipse, a rectangle, and a square, but the second specification shape is smaller than the first specification shape to achieve the purpose of fine division. For the setting of the value of the set number, those skilled in the art can set it by themselves according to the accuracy requirements.

[0135] In an embodiment of the present invention, the method for determining the geometric center of the emphysema region is the same as the method for determining the geometric center of the small airway lesion region, and the method for determining the geometric center of the small airway lesion region is taken as an example for description.

[0136] In an embodiment of the present invention, the method for determining the geometric center of the small airway lesion area is as follows: a plurality of points are set in the small airway lesion area; the distances from the plurality of points to the edge are calculated respectively to obtain a plurality of distances from each point to the edge; and the geometric center of the small airway lesion area is determined according to the plurality of distances from each point to the edge. For example, in the method for determining the geometric center of the emphysema area, a plurality of points are set in the emphysema area, the distances from the plurality of points to the edge are calculated respectively to obtain a plurality of distances from each point to the edge; and the geometric center of the emphysema area is determined according to the plurality of distances from each point to the edge.

[0137] In an embodiment of the present invention, the specific method for determining the geometric center of the small airway lesion area based on the several distances from each point to the edge is: respectively calculating the difference between the several distances from each point to the edge; comparing the difference between the several distances of a certain point with a set value, and if the difference is less than the set value, then accumulating; when the accumulated number is greater than or equal to the preset number, the position of the certain point is the geometric center of the small airway lesion area. The specific method for determining the geometric center of the emphysema area based on the several distances from each point to the edge is the same as the specific method for determining the geometric center of the small airway lesion area based on the several distances from each point to the edge, and reference may be made to the description of the specific method for determining the geometric center of the small airway lesion area based on the several distances from each point to the edge.

[0138] Further, the method for determining the geometric center of the small airway lesion area is taken as an example to illustrate that the method for determining the geometric center of the emphysema area is the same as the method for determining the geometric center of the small airway lesion area: the several points are 10 points, and the positions of the 10 points are within the small airway lesion area. The 10 points are calculated as the center respectively, and are emitted to the edge in the form of several rays (100 rays). When the several rays reach the edge or intersect with the edge, several marks (100 marks) are made; the distance between each point and the several marks is the several distances from each point to the edge.

[0139] Specifically, a point is defined as OX, X represents the nth point, X=1 represents the first point among the 10 points, and a point is emitted to the edge in the form of 100 rays. At this time, the number of markers is 100, namely X1 to X100; the number of distances from a point OX to X1 to X100 on the edge is 100, and the geometric center of the small airway lesion area is determined according to the 100 distances of each of the 10 points.

[0140] In the present invention, the specific method for determining the geometric center of the small airway lesion area based on several distances from each point to the edge is: respectively calculating the difference between several distances from each point to the edge; comparing the difference between the several distances of a certain point with a set value, and if the difference is less than the set value, accumulating; when the accumulated number is greater than or equal to the preset number, the position of the certain point is the geometric center of the small airway lesion area.

[0141] For example, taking one of the several points as an example, 100 rays centered at this point are emitted toward the edge, and the number of distances from this point to the edge is 100. The difference between the 100 is calculated, and each difference is compared with the set value. If the difference is less than the set value, accumulation is performed; when the accumulated number is greater than or equal to the preset number, the position of this point is the geometric center of the small airway lesion area.

[0142] Regarding the setting of the set value and the preset number, those skilled in the art can set them according to the number of distances from the point to the edge and the accuracy required. For example, the set value can be 0.5-1mm; when the number of distances from the point to the edge is 100, the preset number can be 80.

[0143] In an embodiment of the present invention, if the positions of multiple points (such as 2 points) are determined to be the geometric center of the small airway lesion area, the cumulative number of the multiple points is compared, and the one with a larger cumulative number is the geometric center of the small airway lesion area.

[0144] If there are multiple points such as 2 points, namely the first point and the second point, if the positions of the first point and the second point are both determined to be the geometric center of the small airway lesion area, then compare the cumulative number of the first point and the cumulative number of the second point. If the cumulative number of the first point is greater than the cumulative number of the second point, the position of the first point is the geometric center of the small airway lesion area.

[0145] In other embodiments of the present invention, the method for determining the geometric center of the small airway lesion area can also be: shrinking the edge of the small airway lesion area according to a first preset scale to obtain a first edge; selecting a plurality of first reference points on the first edge, and fitting the first reference points to form a first circle; in response to the radius of the first circle being greater than a radius threshold, shrinking the first edge according to a second preset scale until the radius of the circle formed by the plurality of second reference points selected on the shrunken edge is less than the radius threshold; and determining the geometric center of the small airway lesion area using the position mean of the plurality of second reference points or the center of the circle formed by the second reference points. The first preset scale is greater than the second preset scale, such as the first preset scale can be 80% and the second preset scale can be 10%, but this is not a specific limitation of the present invention.

[0146] That is to say, in an embodiment of the present invention, when a small airway lesion area is determined, the edge of the small airway lesion area can be determined to form an edge. Then, the edge can be reduced according to a first preset scale to obtain a first edge. Then, a number of first reference points can be selected from the first edge according to a first rule, wherein the edge of the small airway lesion area can be regarded as consisting of multiple points, and the corresponding first edge can also include multiple points. The first rule can include selecting a number of first reference points according to a preset number of interval points. For example, the preset number of interval points can be 10, and a first reference point can be selected every 10 points on the first edge, wherein the preset number of interval points is not specifically limited. After selecting the first reference point, the first reference point can be curve fitted to form a first circle, and then the radius of the first circle can be determined, wherein the curve fitting method can be the least squares method, but it is not limited to the present invention. If the radius of the first circle is less than or equal to the radius threshold, the center of the first circle can be used as the geometric center of the small airway lesion area, or the mean position of the first reference point can be determined as the geometric center of the small airway lesion area.

[0147] If the radius of the first circle is greater than the radius threshold, it means that the first edge needs to be further reduced to reduce the error. At this time, it can be reduced according to the second preset scale, and the second preset scale is smaller than the first preset scale. When the first edge is reduced to form a second edge (at least one), several second reference points can be selected on the second edge. Similarly, the second reference points can be used to fit a circle (second circle), and it is determined whether the second edge needs to be reduced according to the above rules. If the radius of the second circle fitted by the second reference point is greater than the radius threshold, the second edge is further reduced according to the second preset scale. Conversely, if the radius of the second circle is less than or equal to the radius threshold, the center of the second circle or the position mean of the second reference point can be determined as the geometric center of the small airway lesion area.

[0148] The method for determining the geometric center of the emphysema lesion area is the same as the method for determining the geometric center of the small airway lesion area described above, and will not be repeated here.

[0149] In an embodiment of the present invention, the first grid area corresponding to each adjacent two layers of emphysema, as well as the interlayer spacing and layer thickness are used to obtain the subvolume formed by the adjacent two layers of emphysema, and the first volume of the emphysema region is obtained by the sum of the subvolumes formed by all adjacent two layers of emphysema. Among them, the structure formed by the adjacent two layers of lung region can be regarded as a prism, and the area of ​​the upper and lower bottom surfaces of the prism is the grid area corresponding to the emphysema. The height of the prism can be determined by the interlayer spacing and layer thickness. For example, if the number of layers is N, the height of the prism formed by the first and second layers of emphysema can be the sum of the layer thickness X2 and the interlayer spacing, and the height of the prism formed by the remaining two adjacent layers of emphysema can be the sum of the layer thickness and the interlayer spacing. The subvolume of emphysema formed by the adjacent layers can be determined based on the upper and lower bottom areas and the height. Then, the second volume of emphysema can be obtained by the sum of the subvolumes. That is, since emphysema may exist in multiple places, the first volume of the emphysema region is the total volume of all emphysema.

[0150] In an embodiment of the present invention, the sub-volume formed by the two adjacent layers of small airway lesion regions is obtained by using the second grid area corresponding to each adjacent two layers of small airway lesion regions, as well as the inter-layer spacing and the layer thickness, and the second volume of the small airway lesion region is obtained by using the sum of the sub-volumes formed by all the adjacent two layers of small airway lesion regions. Among them, the structure formed by the adjacent two layers of lung regions can be regarded as a prism, and the area of ​​the upper and lower bottom surfaces of the prism is the grid area corresponding to the small airway lesion region. The height of the prism can be determined by the inter-layer spacing and the layer thickness. For example, if the number of layers is N, the height of the prism formed by the first and second layers of small airway lesion regions can be the sum of the layer thickness*2 and the inter-layer spacing, and the height of the prism formed by the remaining two adjacent layers of small airway lesion regions can be the sum of the layer thickness and the inter-layer spacing. The sub-volume of the small airway lesion region formed by the adjacent layers can be determined based on the upper and lower bottom areas and the height. Then, the second volume of the small airway lesion region can be obtained by using the sum of the sub-volumes. That is, since there may be multiple small airway lesion regions, the second volume of the small airway lesion region is the total volume of all small airway lesion regions.

[0151] In an embodiment of the present invention, the full inspiratory phase lung image and the full expiratory phase lung image can also be used to identify the small airway lesion area of ​​each layer of the image. The method includes: aligning the full inspiratory phase lung image and the full expiratory phase lung image; comparing the aligned full inspiratory phase lung image and the aligned full expiratory phase lung image with the inspiratory phase set threshold and the expiratory phase set threshold respectively; if the CT value of the aligned full inspiratory phase lung image is less than the inspiratory phase set threshold and the aligned full expiratory phase lung image is less than the expiratory phase set threshold, it is considered to be a small airway lesion area; otherwise, it is considered not to be a small airway lesion area.

[0152] In an embodiment of the present invention, the method for registering the full inspiratory phase lung image and the full expiratory phase lung image may use an elastic registration algorithm or utilize the VGG network (VGG-net) in deep learning for registration, or a U network (U-net) for registration. The present invention does not limit the specific registration algorithm.

[0153] In an embodiment of the present invention, the inspiratory phase setting threshold can be set to -950HU, and the expiratory phase setting threshold can be set to -856HU. If the CT value of the registered full inspiratory phase lung image is less than the inspiratory phase setting threshold -950HU, and the CT value of the registered full expiratory phase lung image is less than the expiratory phase setting threshold -856HU, then it is considered that there are small airway lesions in this area; otherwise, it is considered that there are no small airway lesions in this area. In other published papers or in some possible implementations, the inspiratory phase setting threshold and the expiratory phase setting threshold may fluctuate. The present invention does not specifically limit the setting thresholds, and those skilled in the art can appropriately adjust the inspiratory phase setting threshold and the expiratory phase setting threshold.

[0154] In the present invention, the full inspiratory phase lung image or the full expiratory phase lung image can also be used to identify the emphysema area of ​​each layer of the image, including: comparing the full inspiratory phase lung image or the full expiratory phase lung image with the emphysema set threshold; if the CT value of the full inspiratory phase lung image or the full expiratory phase lung image is less than the emphysema set threshold, it is considered to be the emphysema area; otherwise, it is considered not to be the emphysema area.

[0155] In an embodiment of the present invention, the emphysema region is preferably identified by using the full inspiratory phase lung image, and the method for identifying the emphysema region of the full inspiratory phase lung image is: judging whether the CT value of the full inspiratory phase lung image is an emphysema region according to the CT value of the full inspiratory phase lung image and the emphysema setting threshold. Among them, the emphysema setting threshold is a set CT value. Since the CT value of the emphysema region does not change substantially during deep inhalation, and other normal regions (non-emphysema regions) enter the air, the CT value of the air is 1024HU. Since the emphysema region is filled with air, the CT value of the emphysema region is close to 1024HU. In the medical field, the emphysema setting threshold is generally selected as -950HU. If the CT value of the full inspiratory phase lung image is less than -950HU, then this region is judged to be an emphysema region. If the CT value of the full inspiratory phase lung image is greater than or equal to -950HU. Then it is judged that this region is not an emphysema region. In other published papers or in some possible implementations, the set threshold may fluctuate. The present invention does not specifically limit the set threshold, and those skilled in the art may appropriately adjust the set threshold for emphysema.

[0156] In an embodiment of the present invention, the grade classification is a further classification of the lung function grade or the correction grade, so that the current grade is more refined.

[0157] In an embodiment of the present invention, the method for classifying the lung function level according to the first volume and the second volume to obtain the level classification includes: determining the ratio of the first volume and the second volume, classifying the corrected lung function level according to the ratio, and obtaining a first emphysema level category.

[0158] In an embodiment of the present invention, specifically, the lung function level can be classified according to the threshold interval in which the ratio is located. That is, the level category corresponding to the threshold interval in which the ratio is located is determined as the emphysema level is determined as the first emphysema level category. For example, the emphysema level is divided into 3 categories: the first stage or the second stage or the third stage. The ratio threshold interval corresponding to the first stage is [0-0.4], the ratio threshold interval corresponding to the second stage is [0.4-0.8], and the ratio threshold interval corresponding to the first stage is [0.8-1]. When the calculated ratio is 0.7, it is determined that the first emphysema level category is the second stage.

[0159] In the present invention, the method for classifying the lung function level according to the ratio also includes: determining a first threshold and a second threshold; classifying the lung function level according to the ratio, the first threshold and the second threshold, and the lung function level is divided into the first stage, the second stage or the third stage; then, generating a case report according to the lung function data, the emphysema level of the lung function data, and the first stage, the second stage or the third stage. The first threshold and the second threshold can be set to 0.4 and 0.8. The first stage, the second stage or the third stage are respectively the maintainable stage, the dangerous stage and the possible deterioration stage.

[0160] For example, when the emphysema grade is GOLD grade 2 (moderate), the first volume of the emphysema area is divided by the second volume of the small airway lesion area to obtain a ratio, and the ratio is greater than or equal to the second threshold value, indicating that although the emphysema area is at the GOLD grade 2 (moderate), the proportion of the small airway lesion area in the emphysema area is small, and the possibility of the small airway lesion area transforming into the emphysema area is small. The emphysema grade can be determined as the maintainable period of GOLD grade 2 (moderate), and it is considered that drug treatment can be used for improvement.

[0161] After obtaining the above data, a case report can be generated based on the lung function data, emphysema grade and the emphysema grade category, providing comprehensive, reasonable and accurate case report questions to facilitate future treatment of patients.

[0162] In some possible implementations, a diagnosis report may be generated based on the emphysema area, small airway lesion area, emphysema grade, and emphysema grade category obtained corresponding to the lung image. In addition, the lung image may be associated with user information and consultation time, and the user information may include user name, age, contact information, address, and other information related to the user identity. Based on the user information, consultation time, and diagnosis, a case report corresponding to the user information at the consultation time may be generated.

[0163] In other implementations, a patient may have multiple consultations, and an independent case report list may be created for each user information, which may include case reports of all previous consultations.

[0164] Furthermore, the embodiment of the present invention can also compare the case report of the current number of consultations with the previous case reports to obtain a report on the change of the condition. For example, the emphysema grade information, emphysema grade category information, emphysema area, and small airway lesion area in each case report can be obtained by extracting keywords in the case report, so as to obtain at least one of the change information of emphysema grade, change information of emphysema grade category, change information of emphysema area, and change information of small airway lesion area, and a report on the change of the condition can be generated based on the above change information.

[0165] In addition, in the embodiment of the present invention, when the change of emphysema grade information or emphysema grade category change information exceeds the change threshold, a prominent prompt can be given. Since the patient's condition changes quickly or slowly, in the embodiment of the present invention, when the grade changes by more than 2 levels, it is determined to be a fast change. At this time, the patient can be highlighted that the degree of deterioration of the condition is fast or the cure speed is fast. The prominent prompt includes at least one of a voice prompt, a highlight display, and an indicator light prompt.

[0166] In addition, in an embodiment of the present invention, when the rate of change of the emphysema area or the small airway lesion area exceeds a preset ratio, a highlight prompt can be performed. Similarly, excessive changes in the lesion area also need to be taken seriously. In an embodiment of the present invention, the area change rate of the emphysema area or the area change rate of the small airway lesion area in two medical reports can be compared (the area difference divided by the area of ​​the previous lesion area). When the area change rate exceeds 30%, it can be determined that the change is too fast, and a highlight prompt can be given at this time.

[0167] Through the above configuration of the embodiment of the present invention, the automatic detection of emphysema grade can be realized, and the automatic generation of case reports can also be realized, which is more intelligent.

[0168] In addition, an embodiment of the present invention provides a device for detecting emphysema level, such as Figure 5As shown, the device comprises:

[0169] An acquisition module 10, used to acquire image features of a lung image;

[0170] A processing module 20, configured to perform convolution processing on the image features of the lung image to obtain an attention feature matrix;

[0171] The prediction module 30 is used to predict the emphysema grade of the lung image by using the attention feature matrix.

[0172] Optionally, the processing module 20 includes:

[0173] A processing unit, configured to perform a first convolution process on the image features of the lung image to obtain a plurality of reference attention features, wherein the elements in the reference attention features are reference attention points corresponding to the feature points in the image features;

[0174] The processing unit is used to perform a second convolution process on the image feature to obtain an attention coefficient matrix, wherein the elements in the attention coefficient matrix are benchmark attention coefficients corresponding to each of the feature points in the image feature, and the number of channels of the attention coefficient matrix is ​​the same as the number of the benchmark attention features;

[0175] An acquisition unit is used to obtain the attention features of each of the feature points based on the image features, each of the baseline attention features and the attention coefficient matrix, wherein the attention feature matrix is ​​composed of the attention features of each of the feature points.

[0176] Optionally, the acquisition unit is used to obtain a benchmark feature vector corresponding to each benchmark attention feature based on the image feature and each benchmark attention feature, the benchmark feature vector being obtained by the sum of the products of corresponding elements in the image feature and the benchmark attention feature; multiplying each benchmark feature vector with the attention coefficient matrix to obtain a product matrix for each benchmark feature vector; and adding the product values ​​of the same feature point in the product matrix for each benchmark feature vector to obtain the attention feature of each feature point.

[0177] Optionally, the prediction module 30 is used to obtain the emphysema region of the lung image using the attention feature matrix and the region candidate network; and to perform grade prediction on the emphysema region using a classification network to obtain the emphysema grade of the lung image.

[0178] Furthermore, the device also includes:

[0179] a determination module, configured to determine the emphysema grade of the lung function data according to the lung function data of the lung corresponding to the lung image;

[0180] The determination module is further used to determine a first volume of the emphysema region and a second volume of the small airway lesion region according to the lung image;

[0181] a correction module, configured to correct the emphysema level of the lung function data according to the emphysema level of the lung image;

[0182] a classification module, configured to classify the corrected emphysema grade according to the first volume and the second volume to obtain a first emphysema grade category;

[0183] A generating module is used to generate a case report based on the lung function data, the corrected emphysema grade and the first emphysema grade category.

[0184] Furthermore, the device also includes:

[0185] A judgment module, used for judging whether the emphysema level of the lung image is the same as the emphysema level of the lung function data;

[0186] The classification module is further configured to classify the emphysema grade of the pulmonary function data or the emphysema grade of the lung image according to the first volume and the second volume, if they are the same grade, to obtain a second emphysema grade category;

[0187] The generating module is further configured to generate a case report according to the pulmonary function data, the emphysema grade of the pulmonary function data or the emphysema grade of the lung image, and the second emphysema grade category;

[0188] The correction module is specifically used to correct the emphysema level of the lung function data according to the emphysema level of the lung image if the levels are different.

[0189] In addition, an embodiment of the present invention further proposes a computer-readable storage medium, on which a pulmonary emphysema grade detection program is stored. When the pulmonary emphysema grade detection program is executed by a processor, the steps of the pulmonary emphysema grade detection method described below are implemented.

[0190] The various embodiments of the emphysema grade detection device and the computer-readable storage medium of the present invention may refer to the various embodiments of the emphysema grade detection method of the present invention, and will not be described in detail here.

[0191] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0192] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0193] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0194] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for detecting emphysema grade, characterized in that: include: Obtaining image features of a lung image; Performing convolution processing on the image features of the lung image to obtain an attention feature matrix; Predicting the emphysema grade of the lung image using the attention feature matrix; The emphysema grade detection method also includes: Determining the emphysema grade of the lung function data according to the lung function data of the lung corresponding to the lung image; Determine a first volume of the emphysema region and a second volume of the small airway lesion region according to the lung image respectively; Correcting the emphysema level of the lung function data according to the emphysema level of the lung image; classifying the corrected emphysema grade according to the first volume and the second volume to obtain a first emphysema grade category; generating a case report based on the lung function data, the corrected emphysema grade, and the first emphysema grade category; The step of classifying the corrected emphysema grade according to the first volume and the second volume to obtain a first emphysema grade category includes: determining a ratio of the first volume to the second volume; The corrected lung function grade is classified according to the ratio to obtain a first emphysema grade category.

2. The method according to claim 1, characterized in that The step of using the attention feature matrix to predict the emphysema grade of the lung image comprises: Obtaining a pulmonary emphysema region of the lung image using the attention feature matrix and the region candidate network; The classification network is used to predict the grade of the emphysema area to obtain the emphysema grade of the lung image.

3. The method according to claim 1, characterized in that: The performing convolution processing on the image features of the lung image to obtain an attention feature matrix includes: Performing a first convolution process on the image features of the lung image to obtain a plurality of benchmark attention features, wherein the elements in the benchmark attention features are benchmark attentions corresponding to each feature point in the image features; Performing a second convolution process on the image features to obtain an attention coefficient matrix, wherein the elements in the attention coefficient matrix are benchmark attention coefficients corresponding to each of the feature points in the image features, and the number of channels of the attention coefficient matrix is ​​the same as the number of the benchmark attention features; Based on the image features, each of the benchmark attention features and the attention coefficient matrix, the attention features of each of the feature points are obtained, and the attention feature matrix is ​​composed of the attention features of each of the feature points.

4. The method according to claim 3, characterized in that The obtaining of the attention feature of each feature point based on the image feature, each of the benchmark attention features and the attention coefficient matrix comprises: Based on the image feature and each of the benchmark attention features, obtaining a benchmark feature vector corresponding to each benchmark attention feature, wherein the benchmark feature vector is obtained by summing the products of the image feature and corresponding elements in the benchmark attention feature; Multiplying each of the benchmark feature vectors by the attention coefficient matrix to obtain a product matrix for each benchmark feature vector; The product values ​​of the same feature points in the product matrix of each benchmark feature vector are added together to obtain the attention features of each feature point.

5. The method according to claim 1, characterized in that After determining the first volume of the emphysema region and the second volume of the small airway lesion region according to the lung image, the method further includes: Determining whether the emphysema level of the lung image is the same as the emphysema level of the lung function data; If they are the same level, classifying the emphysema level of the pulmonary function data or the emphysema level of the lung image according to the first volume and the second volume to obtain a second emphysema level category; generating a case report according to the lung function data, the emphysema grade of the lung function data or the emphysema grade of the lung image, and the second emphysema grade category; The step of correcting the emphysema level of the lung function data according to the emphysema level of the lung image; If they are different levels, the emphysema level of the lung function data is corrected according to the emphysema level of the lung image.

6. The method according to claim 1, characterized in that Before determining the first volume of the emphysema region and the second volume of the small airway lesion region according to the lung image, the method further includes: Determining the number of slices, scanning slice thickness and slice spacing of the lung image; Determining a first volume of an emphysema region and a second volume of a small airway lesion region according to the lung image comprises: Determine the first volume according to the emphysema area and the number of slices, the scanning slice thickness and the slice spacing; The second volume is determined according to the small airway lesion area, the number of layers, the scanning layer thickness, and the layer spacing.

7. A device for detecting emphysema grade, characterized in that: The device comprises: An acquisition module, used for acquiring image features of a lung image; A processing module, configured to perform convolution processing on the image features of the lung image to obtain an attention feature matrix; A prediction module, configured to predict the emphysema grade of the lung image using the attention feature matrix; A determination module, configured to determine the emphysema grade of the pulmonary function data according to the pulmonary function data of the lung corresponding to the pulmonary image; the determination module is further configured to determine a first volume of the emphysema region and a second volume of the small airway lesion region according to the pulmonary image; a correction module, configured to correct the emphysema level of the lung function data according to the emphysema level of the lung image; A classification module, configured to classify the corrected emphysema grade according to the first volume and the second volume to obtain a first emphysema grade category; the classification module is further configured to determine a ratio of the first volume to the second volume; and classify the corrected lung function grade according to the ratio to obtain a first emphysema grade category; A generating module is used to generate a case report based on the lung function data, the corrected emphysema grade and the first emphysema grade category.

8. A device for detecting emphysema grade, characterized in that: The emphysema level detection device includes: a memory, a processor, and an emphysema level detection program stored in the memory and executable on the processor. When the emphysema level detection program is executed by the processor, the steps of the emphysema level detection method as described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores an emphysema grade detection program, which, when executed by a processor, implements the steps of the emphysema grade detection method according to any one of claims 1 to 6.

Images