Quality control methods, devices, computer equipment and media for pulmonary function test reports

Abstract

The present invention discloses a quality control method, device, computer equipment, and medium for pulmonary function test reports, relating to the field of data processing technology. The method comprises: identifying examination parameter data and graphic data from a PDF file of a pulmonary function test report; extracting coordinate data for each flow rate volume loop from the graphic data; dividing the coordinate data of each flow rate volume loop into different sections based on the coordinate data of the expiratory phase of each flow rate volume loop according to the flow rate variation characteristics and time characteristics; judging whether each flow rate volume loop has an abnormal exhalation condition based on the different sections where the coordinate data is located and whether the coordinate data forms a preset shape feature; and judging whether the examination parameter data of multiple pulmonary function tests in the PDF file meet the result consistency requirements based on the trend differences of the examination parameter data. The present invention is conducive to improving the efficiency and accuracy of quality control of pulmonary function test reports.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a quality control method, device, computer equipment and medium for a pulmonary function test report. Background Art

[0002] Currently, forced vital capacity (FVC) testing is one of the most important tests in pulmonary function testing. The curve collected during FVC testing is called a flow-volume loop. The quality of the FVC loop determines the accuracy of the pulmonary function data calculated from it. During the test, patients may experience abnormal expiratory techniques such as coughing, premature exhalation, insufficient initial exhalation force, hesitant exhalation, and glottal closure. These abnormalities will also be reflected in the FVC loop, which can affect the accuracy of the collected data.

[0003] During actual examinations, especially in community hospitals, pulmonary function tests are not well-versed, requiring higher-level hospitals to remotely interpret reports. Relying on human resources to conduct quality control on a large number of reports from grassroots hospitals cannot meet clinical needs. The aforementioned abnormalities in quality control can only be determined by reviewing the clinical experience and subjective judgment of doctors to determine whether the abnormalities in the pulmonary function flow rate and volume loop are qualified. This makes quality control subjective and inconsistent, which in turn affects accuracy. Summary of the Invention

[0004] In view of this, the present invention provides a quality control method for pulmonary function test reports to solve the technical problems of low efficiency and low accuracy in manual review of reports in the prior art. The method comprises:

[0005] Identify examination parameter data and graphic data from PDF files of pulmonary function test reports;

[0006] For the flow rate capacity rings in the graphic data, extracting coordinate data for each of the flow rate capacity rings;

[0007] Based on the coordinate data of each flow rate volume loop during the exhalation phase, dividing the coordinate data of each flow rate volume loop into a first segment and a second segment according to the change characteristics and time characteristics of the flow rate;

[0008] Determining whether each of the flow rate and capacity loops has an abnormal exhalation condition based on the different sections of the coordinate data and whether the coordinate data forms a preset shape feature, wherein different preset shape features correspond to different abnormal exhalation conditions;

[0009] Based on the trend difference of the examination parameter data, it is determined whether the examination parameter data of multiple pulmonary function tests in the PDF file meet the result consistency.

[0010] Furthermore, based on the coordinate data of each flow rate volume loop during the exhalation phase, the coordinate data of each flow rate volume loop is divided into a first segment and a second segment according to the change characteristics and time characteristics of the flow rate, including:

[0011] determining a maximum flow rate value based on the coordinate data of the expiratory phase of each of the flow rate and volume loops;

[0012] Determine a key time point based on the exhalation time required for the pulmonary function test, and determine the flow rate value corresponding to the key time point, wherein the key time point is the last time point during the exhalation time when the flow rate keeps increasing;

[0013] The minimum value between the maximum flow rate value and the flow rate value corresponding to the key time point is determined as the turning point. According to the generation time sequence of the coordinate data of the exhalation stage, the coordinate data before the turning point is determined as the first segment, and the coordinate data after the turning point is determined as the second segment.

[0014] Furthermore, judging whether each of the flow rate volume loops has an abnormal exhalation condition based on the different sections of the coordinate data and whether the coordinate data forms a preset shape feature includes:

[0015] converting the coordinate data in the second segment from time domain data to frequency domain data, and determining high frequency components in the frequency domain data;

[0016] Constructing a polar coordinate system, mapping the polar angle in the polar coordinate system to vital capacity, mapping the polar diameter in the polar coordinate system to flow velocity, inputting the vital capacity and flow velocity values ​​in the coordinate data corresponding to the high-frequency component into the polar coordinate system, and generating a polar coordinate curve by adjusting the time window length and scaling parameters;

[0017] When the polar diameter of the polar coordinate curve has a first preset shape feature of decreasing and then rising, it is determined that the flow rate capacity loop has the abnormal exhalation condition of coughing.

[0018] Furthermore, judging whether each of the flow rate volume loops has an abnormal exhalation condition based on the different sections where the coordinate data is located and whether the coordinate data forms a preset shape feature includes:

[0019] For the coordinate data in the first section, calculating the gradient of each coordinate data, wherein the gradient of each coordinate data forms a gradient vector;

[0020] constructing a gradient autocorrelation matrix based on the gradient vector;

[0021] Based on the gradient autocorrelation matrix, calculating the eigenvector corresponding to the minimum eigenvalue;

[0022] According to the eigenvector corresponding to the minimum eigenvalue, the point where the gradient direction suddenly changes is determined as a corner point, and at least one corner point is determined;

[0023] A line is formed by connecting the coordinate data corresponding to at least one determined corner point and the coordinate data before and after it. When the line has a first preset shape feature of decreasing and then increasing, it is determined that the flow rate capacity loop has the exhalation abnormality of exhalation initiation hesitation.

[0024] Furthermore, judging whether each of the flow rate and capacity loops has an abnormal exhalation condition based on the different segments of the coordinate data and whether the coordinate data forms a preset shape feature, wherein different preset shape features correspond to different abnormal exhalation conditions, includes:

[0025] Taking the turning point as the center, according to the generation time sequence of the coordinate data, a preset number of continuous coordinate data are taken before and after the turning point to form a local point cloud;

[0026] constructing an α complex or a Vietoris-Rips complex of the local point cloud and performing persistent homology calculation to extract topological features;

[0027] Calculating curvatures of the local point cloud at multiple window size scales with different values ​​to obtain multiple curvatures;

[0028] Calculating the degree of coupling between curvature and topology according to the plurality of curvature and topological features;

[0029] determining, based on the plurality of curvatures, topological features, and coupling degrees between curvature and topology, whether a curve formed by the local point cloud is a second preset shape feature of a flat curve;

[0030] When the curve formed by the local point cloud has a second preset shape feature of a flat curve, and the turning point is smaller than the peak value of the standard flow rate capacity loop, it is determined that the flow rate capacity loop has the exhalation abnormality of insufficient initial exhalation effort.

[0031] Furthermore, calculating the coupling degree between curvature and topology according to the plurality of curvatures and the topological features includes:

[0032] calculating a curvature stability index of the plurality of curvatures;

[0033] constructing a topological feature vector of the topological feature based on the persistent homogeneous death time distribution;

[0034] The product of the curvature stability index and the topological characteristic vector is determined as the coupling degree between the curvature and the topology.

[0035] Furthermore, based on the trend difference of the examination parameter data, determining whether the examination parameter data of multiple pulmonary function tests in the PDF file meet the result consistency includes:

[0036] Calculating the ratio between the examination parameter data of the same examination parameter of every two pulmonary function tests in the multiple pulmonary function tests to obtain the ratios corresponding to the multiple examination parameters;

[0037] Calculate the mean P and standard deviation q of multiple ratios;

[0038] The upper and lower limits of the confidence interval are calculated using the formula P ± 1.96 × q;

[0039] determining a difference range of the inspection parameter data between every two pulmonary function tests according to the maximum and minimum values ​​of the multiple ratios within the confidence interval;

[0040] When the difference range meets the preset threshold range, it is determined that the inspection parameter data of every two pulmonary function tests meet the result consistency.

[0041] The present invention also provides a quality control device for pulmonary function test reports to solve the technical problems of low efficiency and low accuracy in manual review of reports in the prior art. The device includes:

[0042] A data recognition module is used to recognize examination parameter data and graphic data from the PDF file of the pulmonary function test report;

[0043] A coordinate extraction module, configured to extract coordinate data for each flow rate capacity ring in the graphic data;

[0044] a segment division module, configured to divide the coordinate data of each flow rate volume loop into different segments based on the coordinate data of the exhalation phase of each flow rate volume loop and according to the change characteristics and time characteristics of the flow rate;

[0045] an abnormality determination module, configured to determine whether each of the flow rate and capacity loops has an abnormal exhalation condition based on the different sections of the coordinate data and whether the coordinate data forms a preset shape feature, wherein different preset shape features correspond to different abnormal exhalation conditions;

[0046] The consistency judgment module is used to judge whether the examination parameter data of multiple pulmonary function tests in the PDF file meet the result consistency based on the trend difference of the examination parameter data.

[0047] The present invention also provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the quality control method for any of the above-mentioned pulmonary function test reports is implemented to solve the technical problems of low efficiency and low accuracy in manual review of reports in the prior art.

[0048] The present invention also provides a computer-readable storage medium, which stores a computer program for executing any of the above-mentioned quality control methods for pulmonary function test reports, so as to solve the technical problems of low efficiency and low accuracy in manual review of reports in the prior art.

[0049] Compared with the prior art, the beneficial effects of the present invention include: proposing to identify examination parameter data and graphic data from the PDF file of the pulmonary function test report, and extracting coordinate data for each flow rate capacity ring in the graphic data, and then dividing the coordinate data of each flow rate capacity ring into different segments according to the flow rate change characteristics and time characteristics, judging whether each flow rate capacity ring has abnormal exhalation according to the different segments where the coordinate data is located and whether the coordinate data forms a preset shape feature, judging whether the examination parameter data of multiple pulmonary function tests in the PDF file meet the result consistency based on the trend difference of the examination parameter data, realizing automatic and intelligent abnormality detection and quality control of the examination parameter data and graphic data of the pulmonary function test report, compared with the manual review in the prior art, the pulmonary function test report can be interpreted more efficiently, which is conducive to meeting the quality control needs of a large number of reports; in addition, the standardization of abnormality detection and quality control of pulmonary function test reports is realized, avoiding the influence of individual differences, and is conducive to improving the accuracy of report quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0051] Figure 1 1 is a schematic diagram of a flow rate capacity loop curve provided by an embodiment of the present invention;

[0052] Figure 2 This is a flow chart of a quality control method for a pulmonary function test report provided by an embodiment of the present invention;

[0053] Figure 3 is a schematic diagram of a pulmonary function test report provided by an embodiment of the present invention;

[0054] Figure 4 Schematic diagram of a flow rate volume loop curve with different abnormal exhalation conditions provided by an embodiment of the present invention;

[0055] Figure 5 This is a structural block diagram of a computer device provided by an embodiment of the present invention;

[0056] Figure 6 This is a structural block diagram of a quality control device for a pulmonary function test report provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0058] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that, in the absence of conflict, the features in the following embodiments and embodiments can be combined with each other. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative work are within the scope of protection of this application.

[0059] In an embodiment of the present invention, an example of a flow rate volume loop curve collected by forced vital capacity testing is as follows: Figure 1 As shown, 1, 2, and 3 represent flow rate capacity loops formed by three forced vital capacity tests, respectively, B represents the best flow rate capacity loop obtained from the three forced vital capacity tests, the X-axis represents vital capacity, and the Y-axis represents flow rate.

[0060] In an embodiment of the present invention, a quality control method for a pulmonary function test report is provided, such as Figure 2 As shown, the method includes:

[0061] Step S201: identifying examination parameter data and graphic data from the PDF file of the pulmonary function examination report;

[0062] Step S202: extracting coordinate data for each flow rate capacity ring in the graphic data;

[0063] Step S203: Based on the exhalation phase of each flow volume loop (such as Figure 1As shown, the curve stage above the X-axis in the flow rate capacity loop is the exhalation stage) coordinate data, and the coordinate data of each flow rate capacity loop is divided into different segments according to the change characteristics and time characteristics of the flow rate;

[0064] Step S204: determining whether each of the flow rate and volume loops has an abnormal exhalation condition based on the different sections of the coordinate data and whether the coordinate data forms a preset shape feature, wherein different preset shape features correspond to different abnormal exhalation conditions;

[0065] Step S205: Based on the trend difference of the examination parameter data, it is determined whether the examination parameter data of multiple pulmonary function tests in the PDF file meet the result consistency.

[0066] Depend on Figure 2 As can be seen from the process shown, in an embodiment of the present invention, automatic and intelligent abnormality detection and quality control of the examination parameter data and graphic data of the pulmonary function test report are realized. Compared with the manual review in the prior art, the pulmonary function test report can be interpreted more efficiently, which is conducive to meeting the demand for quality control of a large number of reports; in addition, the standardization of abnormality detection and quality control of pulmonary function test reports is also realized, which avoids the influence of individual differences and is conducive to improving the accuracy of report quality control.

[0067] In specific implementation, in the process of identifying the examination parameter data and graphic data from the PDF file of the pulmonary function examination report, the PDF format pulmonary function examination report can be divided into basic information, examination parameter data and graphic data, such as Figure 3 The patient's basic information is as shown in Figure 3 The top part includes name, age, examination department, etc., and the examination parameter data is as follows: Figure 3 As shown in the middle part of the diagram, it includes data corresponding to examination parameters such as FVC and FEV, and graphic data such as Figure 3 As shown in the bottom part, it includes curves such as flow rate capacity loop, and then the inspection parameter data and graphic data are identified for different parts through OCR text and graphic recognition.

[0068] In a specific implementation, in the process of extracting coordinate data for each of the flow rate and capacity loops, different flow rate and capacity loop curves are separated according to the different colors of the flow rate and capacity loop curves in the pulmonary function test report (for example, the original image color of the pixels in the flow rate and capacity loop curve is restored, and the image is corroded to expand the color characteristics of the curve; the image after corrosion is obtained is divided, and the number of occurrences of pixel values ​​of different colors is counted. When the number of occurrences of a color reaches a set threshold, it is considered that the curve of the entire image contains this category of color; the obtained color RGB values ​​are screened, similar colors are deleted, and only one main color is retained. The main color is used as a basis to extract pixel blocks of this color in the image to obtain a curve corresponding to the main color). Then, coordinate data is extracted for each flow rate and capacity loop curve. The horizontal coordinate data in the coordinate data is the vital capacity, and the vertical coordinate data is the flow rate. The horizontal coordinate data of each flow rate and capacity loop curve can be formed into a first array in the order of generation time, and the vertical coordinate data of each flow rate and capacity loop curve can be formed into a second array in the order of generation time.

[0069] In specific implementation, the present application does not specifically limit the specific method for extracting the horizontal coordinate data and the vertical coordinate data of each flow rate capacity loop curve. For example, based on the vertical relationship between the X coordinate axis and the Y coordinate axis, the pixel coordinates of the X coordinate axis and the Y coordinate axis can be extracted based on the curve image, and then the pixel coordinates can be converted into real coordinates by the following formulas (1) and (2) to obtain coordinate data:

[0070]

[0071] Among them, X, Y represent the coordinate values ​​of the actual curve point data required for the final conversion; pX, pY are the pixel-level coordinates of each point on the curve; X min and Y min The data values ​​of the first detected digital coordinate data on the X-axis and Y-axis starting from the origin respectively; when obtaining the character values ​​on the coordinate axis, character detection and recognition are required. At the same time, the character detection will draw a rectangular detection box for each digital character, pX i , pX j Represents the horizontal coordinate of the upper left corner vertex pixel of the rectangular box during the numerical detection of the two coordinate axes; X i , X j is the specific value of coordinate detection; pY i , pY j Represents the vertical coordinate of the pixel at the upper left corner of the rectangular box during the detection of the upper and lower coordinate axes of the Y axis; i , Y j The specific value after character recognition for coordinate detection.

[0072] In specific implementation, in order to accurately and efficiently determine whether there is an abnormal exhalation condition in the flow rate volume loop based on the coordinate data, it is proposed to divide the coordinate data of the exhalation phase of each flow rate volume loop into different segments according to the change characteristics and time characteristics of the flow rate, and then determine whether there is an abnormal exhalation condition in each flow rate volume loop based on the differences in the segments and preset shape characteristics.

[0073] For example, according to the change characteristics and time characteristics of the flow rate, the process of dividing the coordinate data of the exhalation phase of each flow rate volume loop into different segments includes the following steps:

[0074] Determining the maximum value of the flow rate value based on the coordinate data of the expiratory phase of each flow rate volume loop (i.e., determining the maximum value among the vertical coordinate data of a certain flow rate volume loop);

[0075] Determine a key time point based on the exhalation duration required for the pulmonary function test, and determine the flow rate value corresponding to the key time point. The key time point is the last time point during the exhalation duration at which the flow rate continues to increase (the key time point can be determined based on the curve trend of the flow rate volume loop during the exhalation process. For example, for a conventional pulmonary function test, the exhalation duration is generally 1 second, which is the last time point at which the flow rate continues to increase. The key time point can be 1 second).

[0076] The minimum value of the maximum flow rate value and the flow rate value corresponding to the key time point is determined as the turning point. According to the generation time sequence of the coordinate data in the exhalation stage, the coordinate data before the turning point is determined as the first segment, and the coordinate data after the turning point is determined as the second segment. Specifically, the first segment and the second segment are divided by the maximum value of the flow rate value and the minimum value of the flow rate value corresponding to the key time point, which can effectively avoid the cross-border problem.

[0077] During specific implementation, in the process of determining the flow rate value corresponding to the key time point, in order to adapt to individual differences and improve accuracy, the flow rate value corresponding to the key time point can be a dynamic value determined dynamically. For example, the average of the flow rate values ​​before the key time point or a preset proportion of the average can be determined as the flow rate value corresponding to the key time point.

[0078] In specific implementation, in order to efficiently and accurately determine whether there is an abnormal exhalation condition in each of the flow rate capacity loops, the coughing condition can be detected by detecting whether there is a curve trend in which the flow rate first decreases and then increases. For example, if the subsequent flow rate value in multiple consecutive flow rate values ​​is greater than the previous flow rate value, it means that the curve has an increasing flow rate trend; if the subsequent flow rate value in multiple consecutive flow rate values ​​is less than the previous flow rate value, it means that the curve has a decreasing flow rate trend; if the flow rate value shows a trend of first decreasing and then increasing, it is determined that there is an abnormal coughing condition.

[0079] In addition, considering that abnormal exhalation due to coughing generally occurs in the second segment and the coordinate data of the second segment is relatively large, in order to quickly and accurately determine whether there is abnormal exhalation due to coughing, the following cough detection method is proposed:

[0080] The coordinate data in the second segment is converted from time domain data to frequency domain data, and the high-frequency component is determined in the frequency domain data. Specifically, the coordinate data can be converted from time domain data to frequency domain data using the Fourier transform method (the spectrum resolution of the Fourier transform is determined by the sampling rate and the window length. The longer the window length, the higher the spectrum resolution, but it may cause the loss of time domain details. Therefore, the specific value of the window length is determined according to the actual demand for spectrum resolution). Then, the high-frequency component part of the frequency domain data can be determined by analyzing the high-frequency energy ratio of the high-frequency component (for example, the high-frequency frequency range can be determined in the spectrum, For example, a frequency band above the main frequency of the signal is determined as a high-frequency frequency range, and the energy of each frequency point within the high-frequency frequency range is calculated. The high-frequency energy ratio is calculated based on the energy of each frequency point. If the high-frequency energy ratio is greater than a preset threshold, the frequency point within the high-frequency frequency range is determined as a high-frequency component. The preset threshold can be a fixed value, such as 5% to 10%, or a dynamic threshold. For example, an energy accumulation curve is plotted, and the frequency corresponding to the slope mutation point in the energy accumulation curve is used as the threshold. The high-frequency component portion indicates that a high-frequency change or a sawtooth shape has occurred in the curve. This method can roughly and quickly determine which part or area of ​​the curve is likely to have a cough.

[0081] Constructing a polar coordinate system, mapping the polar angle in the polar coordinate system to vital capacity, mapping the polar diameter in the polar coordinate system to flow velocity, inputting the vital capacity and flow velocity values ​​in the coordinate data corresponding to the high-frequency component into the polar coordinate system, respectively, and generating a polar coordinate curve by adjusting the time window length (a shorter time window length can be used to achieve high time resolution to capture rapid changes, and the specific data of the time window length can be determined according to specific needs) and scaling parameters (such as amplitude scaling and / or angle scaling, amplitude scaling is used to control the display range of the polar coordinate radius, such as normalization or logarithmic scaling, and angle scaling is used to adjust the mapping ratio of the angle, and the specific adjustment ratios of amplitude scaling and angle scaling can be determined according to specific needs);

[0082] If the polar diameter of the polar coordinate curve has a first preset shape feature of decreasing and then rising, it is determined that the flow rate capacity loop has the abnormal exhalation condition of coughing.

[0083] That is, by using the high-frequency component, it is roughly and quickly determined which part or area of ​​the curve is most likely to have an abnormal cough, and then a polar coordinate curve is generated for the coordinate data corresponding to the high-frequency component by using the polar coordinate method. If the polar diameter of the polar coordinate curve decreases and then increases, it means that the part of the coordinate data corresponding to the flow rate capacity loop curve has a first preset shape feature of first decreasing and then increasing, that is, there is an abnormal cough exhalation, such as Figure 4 The situation shown by the circled curve in Figure (a).

[0084] In specific implementation, considering that abnormal exhalation with initial exhalation hesitation generally occurs in the first segment and the coordinate data of the first segment is relatively small, in order to quickly and accurately determine whether there is abnormal exhalation with initial exhalation hesitation, the following method for detecting initial exhalation hesitation is proposed:

[0085] For the coordinate data in the first section, calculate the gradient of each coordinate data (for example, the gradient of the coordinate data may be calculated using a difference method), and the gradient of each coordinate data forms a gradient vector;

[0086] constructing a gradient autocorrelation matrix based on the gradient vector;

[0087] Based on the gradient autocorrelation matrix, calculating the eigenvector corresponding to the minimum eigenvalue;

[0088] According to the eigenvector corresponding to the minimum eigenvalue, points where the gradient direction suddenly changes are identified as corner points (for example, points where the gradient direction suddenly changes can be determined by non-maximum suppression and threshold screening, that is, calculating the corner point response value R of each point, setting the corner point response threshold T, screening points with R>T as candidate corner points, and using non-maximum suppression, in a neighborhood window centered on the current point, selecting the maximum point among the candidate corner points in a preset neighborhood as the corner point) to determine at least one corner point;

[0089] The coordinate data corresponding to at least one determined corner point is connected with the coordinate data before and after it to form a line. If the line has a first preset shape feature of decreasing and then rising (i.e., the first preset shape feature of decreasing and then rising appears on the flow rate capacity loop curve, such as Figure 4 If the flow rate volume loop has the abnormal exhalation condition of exhalation initiation hesitation, as shown in the circled curve in the middle figure (d), it is determined that the flow rate volume loop has the abnormal exhalation condition of exhalation initiation hesitation.

[0090] In specific implementation, the existing gradient calculation method can be used to calculate the gradient of the coordinate data. For example, the central difference quotient can be used to approximately calculate the gradient, including the following steps:

[0091] Step 1. Input discrete coordinate data: {(x1,y1),(x2,y2),...,(x N ,y N )};

[0092] Step 2: Set the spacing h:

[0093] h=x i+1 -x i ,

[0094] In the formula, h is the distance between points, x is i is the horizontal coordinate value of the i-th point, x i+1 is the horizontal coordinate value of the i+1th point;

[0095] Step 3: Calculate the gradient of each intermediate point:

[0096]

[0097] Where, gradient i is the gradient of the i-th point, y i+1 and y i-1 are the ordinate values ​​of the i+1th and i-1th points respectively, and N is the total number of points in the coordinate data;

[0098] Step 4. Calculate the gradient at the boundary points using forward and backward differences:

[0099] Forward difference (i=1):

[0100]

[0101] Backward difference (i=N):

[0102]

[0103] In the formula, gradient1 is the gradient of the first coordinate point, y1 and y2 are the ordinate values ​​of the first and second coordinate points respectively, gradient N is the gradient of the last coordinate point, y N and y N-1 The ordinate values ​​of the Nth and N-1th points respectively;

[0104] Step 5. Arrange the gradients of all points in order to form a gradient vector:

[0105]

[0106] Where g is the gradient vector of the coordinate data.

[0107] In specific implementation, the corner response value R can be calculated according to the Harris corner response function formula: R = λ1λ2-k(λ1+λ2) 2 =det(M)-k·trace 2 (M),

[0108]

[0109] Where R is the corner response function value, which indicates the corner possibility strength; λ1 and λ2 are the eigenvalues ​​of the structure matrix M, which describe the gradient change; k is an empirical constant, generally 0.04 to 0.06; det(M) is the determinant of the matrix M, trace(M) is the trace of the matrix M, that is, the sum of the diagonal elements; M is the structure tensor, that is, the Harris matrix; I x and I y are the gradients of the coordinate data in the x and y directions respectively.

[0110] In practice, to quickly and accurately determine whether there is an abnormal exhalation condition characterized by insufficient initial exhalation effort, we propose the following method, which combines persistent coherent topological features with curve geometric features (using persistent coherent features to capture global structure and geometric features such as curvature to analyze local changes), to detect insufficient initial exhalation effort. This method achieves highly robust judgment:

[0111] Taking the turning point as the center, and in the order of the generation time of the coordinate data, taking a preset number (for example, the preset number can be 10, 20, etc.) of continuous coordinate data before and after the turning point to form a local point cloud (for example, first, converting the preset number of the above-mentioned two-dimensional coordinate data into three-dimensional coordinates (x, y, z) to obtain multiple point cloud data, for example, assuming that all points are located in the same plane, that is, z = 0, or obtaining the z value through a monocular depth estimation algorithm (such as MonoDepth) or binocular disparity calculation; then, importing the multiple point cloud data into ContextCapture software, the software can automatically perform multiple key steps including data preprocessing, feature extraction and final model generation, and output a point cloud three-dimensional model);

[0112] Constructing an α-complex or a Vietoris-Rips complex of the local point cloud and performing persistent homology calculations to extract topological features (e.g., extracting topological features from the α-complex or the Vietoris-Rips complex using the GUDHI TDA-tutorial, where the topological features may include 0-persistent barcodes and 1-persistent barcodes);

[0113] The curvature of the local point cloud under multiple different window size scales is calculated (for example, under a certain window size scale, the point cloud data under the window is obtained and a surface is formed, and the surface normal vector is calculated for a point F on the surface, such as the partial derivatives of the x, y, and z axes of the point F(x, y, z) are respectively calculated to obtain F x 、F y 、F z , substitute these partial derivatives into the vector (F x ,F y ,F z ), we get the normal vector); then we can calculate the curvature σ at a point on the surface by the following formula: β0 is the change of the surface along the normal vector, β1 and β2 are the distribution of this point on the tangent plane), and multiple curvatures are obtained;

[0114] Calculating a degree of coupling between curvature and topology according to the plurality of curvatures and the topological features;

[0115] determining, based on the plurality of curvatures, topological features, and coupling degrees between curvature and topology, whether a curve formed by the local point cloud is a second preset shape feature of a flat curve;

[0116] If (if the flow rate capacity loop curve has a flat curve, that is, there is a second preset shape feature, such as Figure 4 If the turning point is smaller than the peak value of the standard flow rate capacity loop (which may be the standard flow rate capacity loop under normal circumstances), it is determined that the flow rate capacity loop has the abnormal exhalation condition of insufficient initial exhalation effort.

[0117] In a specific implementation, calculating the coupling degree between curvature and topology according to the multiple curvatures and the topological features includes:

[0118] calculating a curvature stability index of the plurality of curvatures;

[0119] constructing a topological feature vector of the topological feature based on the persistent homogeneous death time distribution;

[0120] The product of the curvature stability index and the topological characteristic vector is determined as the coupling degree between the curvature and the topology.

[0121] In specific implementations, after obtaining multiple curvatures, topological features, and the degree of coupling between curvature and topology, a trained neural network model can be used to determine whether the curve formed by the local point cloud is a flat curve. For example, the relevant geometric features of multiple curvatures (such as mean curvature, curvature variance, first-order derivative, etc.), topological features (such as the number of 0-permanent barcodes, the number of 1-permanent barcodes, topological feature vectors, etc.), and the degree of coupling between curvature and topology can be input into the trained neural network model, and the trained neural network model outputs a character or mark indicating whether it is a flat curve.

[0122] In specific implementation, the curvature stability index can be calculated by the following formula:

[0123]

[0124] in, is the curvature stability index, k is the average curvature of multiple curvatures, and avg() is the aggregation function.

[0125] In specific implementation, the topological eigenvector can be calculated using the following formula:

[0126]

[0127] Among them, f is the topological eigenvector, m r is the death time of the rth persistent barcode, n r is the birth time of the rth persistent barcode, and Barcode1 is the total number of persistent barcodes.

[0128] In specific implementation, in order to quickly and accurately determine whether there is an abnormal exhalation condition such as premature exhalation termination, the following method for detecting premature exhalation termination is proposed:

[0129] Take multiple coordinate data at the end of the second segment and analyze the slope of the multiple coordinate data. If the slope is continuously low, it is determined that there is an abnormal exhalation condition in which exhalation is terminated prematurely, such as Figure 4 The situation shown by the circled curve on the right side of the middle figure (c).

[0130] During specific implementation, the expiratory volume can also be verified. If the expiratory volume is ≤5% FVC or 150 ml, the start of exhalation is determined to be valid.

[0131] In specific implementation, the consistency of the results of multiple test data can be verified by the FVC difference. For example, the two largest FVCs in multiple tests are taken and the difference between the two largest FVCs is calculated. If the difference is ≤150ml, it is judged that the multiple test data meet the result consistency.

[0132] In addition, the measurement system variability can be quantified to accurately verify the consistency of the results of multiple test data. For example, the ratio between the test parameter data of the same test parameter (such as FVC, FEV, PEF, etc.) of each two pulmonary function tests in multiple pulmonary function tests can be calculated to obtain the ratios corresponding to multiple test parameters.

[0133] Calculate the mean P and standard deviation q of multiple ratios;

[0134] The upper limit (i.e., P+1.96×q is the upper limit) and lower limit (i.e., P-1.96×q is the lower limit) of the confidence interval are calculated using the formula P±1.96×q;

[0135] determining a difference range of the examination parameter data between every two pulmonary function tests based on the maximum and minimum values ​​of the multiple ratios within the confidence interval (i.e., for the data of a certain examination parameter, finding the maximum and minimum values ​​from the multiple ratios falling within the confidence interval, and taking the difference between the maximum and minimum values ​​as the difference range);

[0136] If the difference range meets the preset threshold range, it is determined that the inspection parameter data of every two pulmonary function tests meet the result consistency.

[0137] During specific implementation, the qualification of the report can be determined based on whether there are abnormal exhalations in each flow rate volume loop and whether the inspection parameter data of multiple pulmonary function tests meet the result consistency. If the report is unqualified due to abnormal exhalations and / or the inspection parameter data does not meet the result consistency, an information reminder will be pushed to the user, indicating that the patient's report has unqualified report characteristics and needs to be re-examined, and the doctor will be prompted on how to direct the patient's examination to avoid such situations.

[0138] In this embodiment, a computer device is provided, such as Figure 5 As shown, it includes a memory 501, a processor 502 and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, any of the above-mentioned quality control methods for pulmonary function test reports is implemented.

[0139] Specifically, the computer device may be a computer terminal, a server or a similar computing device.

[0140] In this embodiment, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program for executing any of the above-mentioned quality control methods for pulmonary function test reports.

[0141] Specifically, computer-readable storage media include permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable storage media does not include transitory media such as modulated data signals and carrier waves.

[0142] Based on the same inventive concept, an embodiment of the present invention further provides a quality control device for a pulmonary function test report, as described in the following embodiments. Since the principle of solving the problem by the quality control device for a pulmonary function test report is similar to that of the quality control method for a pulmonary function test report, the implementation of the quality control device for a pulmonary function test report can refer to the implementation of the quality control method for a pulmonary function test report, and the repeated parts will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and conceived.

[0143] Figure 6 This is a structural block diagram of a quality control device for a pulmonary function test report according to an embodiment of the present invention. Figure 6 Shown, including:

[0144] The data identification module 601 is used to identify the examination parameter data and graphic data from the PDF file of the pulmonary function examination report;

[0145] A coordinate extraction module 602 is configured to extract coordinate data for each flow rate capacity ring in the graphic data;

[0146] a segment division module 603 for dividing the coordinate data of each flow rate volume loop into different segments based on the coordinate data of the exhalation phase of each flow rate volume loop and according to the change characteristics and time characteristics of the flow rate;

[0147] An abnormality determination module 604 is configured to determine whether each of the flow rate and volume loops has an abnormal exhalation condition based on the different segments of the coordinate data and whether the coordinate data forms a preset shape feature, where different preset shape features correspond to different abnormal exhalation conditions;

[0148] The consistency judgment module 605 is configured to judge whether the examination parameter data of multiple pulmonary function tests in the PDF file meet result consistency based on trend differences of the examination parameter data.

[0149] In one embodiment, a segment division module is used to determine the maximum flow rate value based on the coordinate data of the expiratory phase of each flow rate capacity loop; determine the key time point according to the exhalation duration required for the pulmonary function test, and determine the flow rate value corresponding to the key time point, wherein the key time point is the last time point in the exhalation duration at which the flow rate keeps increasing; determine the minimum value between the maximum flow rate value and the flow rate value corresponding to the key time point as the turning point, and determine the coordinate data before the turning point as the first segment and the coordinate data after the turning point as the second segment according to the generation time sequence of the coordinate data of the expiratory phase.

[0150] In one embodiment, the abnormality judgment module is used to convert the coordinate data in the second segment from time domain data to frequency domain data, and determine the high-frequency component in the frequency domain data; construct a polar coordinate system, map the polar angle in the polar coordinate system to vital capacity, map the polar diameter in the polar coordinate system to flow velocity, input the vital capacity and flow velocity values ​​in the coordinate data corresponding to the high-frequency component into the polar coordinate system respectively, and generate a polar coordinate curve by adjusting the time window length and scaling parameters; if the polar diameter of the polar coordinate curve has a first preset shape feature of decreasing and then rising, it is judged that the flow rate capacity loop has the exhalation abnormality of coughing.

[0151] In one embodiment, the abnormality judgment module is used to calculate the gradient of each coordinate data within the coordinate data in the first section, and the gradient of each coordinate data forms a gradient vector; construct a gradient autocorrelation matrix based on the gradient vector; calculate the eigenvector corresponding to the minimum eigenvalue based on the gradient autocorrelation matrix; determine the point where the gradient direction suddenly changes as a corner point according to the eigenvector corresponding to the minimum eigenvalue, and determine at least one corner point; form a line between the coordinate data corresponding to the at least one corner point and the coordinate data before and after it, and if the line has a first preset shape feature of decreasing and then rising, it is judged that the flow rate capacity loop has the exhalation abnormality of hesitating at the beginning of exhalation.

[0152] In one embodiment, an abnormality judgment module is used to take a preset number of continuous coordinate data before and after the turning point, with the turning point as the center, in the order of the generation time of the coordinate data, to form a local point cloud; construct an α complex or Vietoris-Rips complex of the local point cloud and perform persistent homology calculation to extract topological features; calculate the curvature of the local point cloud under multiple window size scales with different window sizes to obtain multiple curvatures; calculate the coupling degree of curvature and topology based on the multiple curvatures and the topological features; determine whether the curve formed by the local point cloud is a second preset shape feature of a flat curve based on the multiple curvatures, the topological features and the coupling degree of the curvature and topology; if so, and the turning point is smaller than the peak value of the standard flow rate capacity loop, it is determined that the flow rate capacity loop has the exhalation abnormality of insufficient exhalation starting force.

[0153] In one embodiment, the abnormality judgment module is used to calculate the curvature stability index of the multiple curvatures; construct a topological feature vector of the topological feature based on the continuous homophonic death time distribution; and determine the product of the curvature stability index and the topological feature vector as the coupling degree between the curvature and the topology.

[0154] In one embodiment, a consistency judgment module is used to calculate the ratio between the inspection parameter data of the same inspection parameter of each two pulmonary function tests in multiple pulmonary function tests to obtain ratios corresponding to multiple inspection parameters; calculate the mean P and standard deviation q of multiple ratios; calculate the upper and lower limits of the confidence interval by the formula P±1.96×q; determine the difference range of the inspection parameter data of each two pulmonary function tests based on the maximum and minimum values ​​of the multiple ratios within the confidence interval; if the difference range meets the preset threshold range, it is judged that the inspection parameter data of each two pulmonary function tests meet the result consistency.

[0155] The embodiments of the present invention achieve the following technical effects: automatic and intelligent abnormality detection and quality control of the examination parameter data and graphic data of the pulmonary function test report are realized. Compared with the manual review in the prior art, the pulmonary function test report can be interpreted more efficiently, which is conducive to meeting the needs of quality control of a large number of reports; in addition, the standardization of abnormality detection and quality control of pulmonary function test reports is also achieved, avoiding the impact of individual differences, which is conducive to improving the accuracy of report quality control.

[0156] Obviously, those skilled in the art should understand that the various modules or steps of the above-mentioned embodiments of the present invention can be implemented using a general-purpose computing device, they can be concentrated on a single computing device, or distributed across a network composed of multiple computing devices. Alternatively, they can be implemented using program code executable by the computing device, so that they can be stored in a storage device and executed by the computing device. In some cases, the steps shown or described can be performed in a different order than herein, or they can be made into separate integrated circuit modules, or multiple modules or steps can be made into a single integrated circuit module for implementation. Thus, the embodiments of the present invention are not limited to any specific combination of hardware and software.

[0157] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A quality control method for a pulmonary function test report, characterized in that: include: Identify examination parameter data and graphic data from PDF files of pulmonary function test reports; For the flow rate capacity rings in the graphic data, extracting coordinate data for each of the flow rate capacity rings; Based on the coordinate data of the expiratory phase of each flow rate volume loop, the coordinate data of the expiratory phase of each flow rate volume loop is divided into a first segment and a second segment according to the change characteristics and time characteristics of the flow rate; Based on the different segments of the coordinate data and whether the coordinate data form a preset shape feature, determining whether each of the flow rate and capacity loops has an abnormal exhalation condition, wherein different preset shape features correspond to different abnormal exhalation conditions, wherein the abnormal exhalation condition of insufficient initial exhalation effort is detected by combining the persistent coherent topological feature and the curve geometric feature; Based on the trend difference of the examination parameter data, determining whether the examination parameter data of multiple pulmonary function tests in the PDF file meet the result consistency; Based on the coordinate data of the expiratory phase of each flow rate volume loop, the coordinate data of the expiratory phase of each flow rate volume loop is divided into a first segment and a second segment according to the change characteristics and time characteristics of the flow rate, including: determining a maximum flow rate value based on the coordinate data of the expiratory phase of each of the flow rate and volume loops; Determine a key time point based on the exhalation time required for the pulmonary function test, and determine the flow rate value corresponding to the key time point, wherein the key time point is the last time point during the exhalation time when the flow rate keeps increasing; Determine the minimum value of the maximum flow velocity value and the flow velocity value corresponding to the key time point as the turning point, and determine the coordinate data before the turning point as the first segment and the coordinate data after the turning point as the second segment based on the generation time sequence of the coordinate data of the exhalation phase; Judging whether each of the flow rate volume loops has an abnormal exhalation condition according to different sections of the coordinate data and whether the coordinate data forms a preset shape feature includes: converting the coordinate data in the second segment from time domain data to frequency domain data, and determining high frequency components in the frequency domain data; Constructing a polar coordinate system, mapping the polar angle in the polar coordinate system to vital capacity, mapping the polar diameter in the polar coordinate system to flow velocity, inputting the vital capacity and flow velocity values ​​in the coordinate data corresponding to the high-frequency component into the polar coordinate system, and generating a polar coordinate curve by adjusting the time window length and scaling parameters; When the polar diameter of the polar coordinate curve has a first preset shape characteristic of decreasing and then rising, it is determined that the flow rate capacity loop has the abnormal exhalation condition of coughing; Determining whether each of the flow rate volume loops has an abnormal exhalation condition according to different sections of the coordinate data and whether the coordinate data forms a preset shape feature includes: For the coordinate data in the first section, calculating the gradient of each coordinate data, wherein the gradient of each coordinate data forms a gradient vector; constructing a gradient autocorrelation matrix based on the gradient vector; Based on the gradient autocorrelation matrix, calculating the eigenvector corresponding to the minimum eigenvalue; According to the eigenvector corresponding to the minimum eigenvalue, the point where the gradient direction suddenly changes is determined as a corner point, and at least one corner point is determined; A line is formed by connecting the coordinate data corresponding to at least one determined corner point and the coordinate data before and after it. When the line has a first preset shape feature of decreasing and then increasing, it is determined that the flow rate capacity loop has the exhalation abnormality of exhalation initiation hesitation.

2. The quality control method for a pulmonary function test report according to claim 1, wherein: According to different sections of the coordinate data and whether the coordinate data forms a preset shape feature, it is determined whether each of the flow rate capacity loops has an abnormal exhalation condition, and different preset shape features correspond to different abnormal exhalation conditions, including: Taking the turning point as the center, according to the generation time sequence of the coordinate data, a preset number of continuous coordinate data are taken before and after the turning point to form a local point cloud; constructing an α complex or a Vietoris-Rips complex of the local point cloud and performing persistent homology calculation to extract topological features; Calculating curvatures of the local point cloud at multiple window size scales with different values ​​to obtain multiple curvatures; Calculating the degree of coupling between curvature and topology according to the plurality of curvature and topological features; determining, based on the plurality of curvatures, topological features, and coupling degrees between curvature and topology, whether a curve formed by the local point cloud is a second preset shape feature of a flat curve; When the curve formed by the local point cloud has a second preset shape feature of a flat curve, and the turning point is smaller than the peak value of the standard flow rate capacity loop, it is determined that the flow rate capacity loop has the exhalation abnormality of insufficient initial exhalation effort.

3. The quality control method for a pulmonary function test report according to claim 2, wherein: Calculating a degree of coupling between curvature and topology according to the plurality of curvatures and the topological features, including: calculating a curvature stability index for the plurality of curvatures; constructing a topological feature vector of the topological feature based on the persistent homogeneous death time distribution; The product of the curvature stability index and the topological eigenvector is determined as the coupling degree between the curvature and the topology.

4. The quality control method for pulmonary function test report according to claim 1, characterized in that: Based on the trend difference of the examination parameter data, determining whether the examination parameter data of multiple pulmonary function tests in the PDF file meet the result consistency includes: Calculating the ratio between the examination parameter data of the same examination parameter of every two pulmonary function tests in the multiple pulmonary function tests to obtain the ratios corresponding to the multiple examination parameters; Calculate the mean P and standard deviation q of multiple ratios; The upper and lower limits of the confidence interval are calculated using the formula P ± 1.96 × q; determining a difference range of the inspection parameter data between every two pulmonary function tests according to the maximum and minimum values ​​of the multiple ratios within the confidence interval; When the difference range meets the preset threshold range, it is determined that the inspection parameter data of every two pulmonary function tests meet the result consistency.

5. A quality control device for executing the quality control method for a pulmonary function test report according to claim 1, characterized in that: include: A data recognition module is used to recognize examination parameter data and graphic data from the PDF file of the pulmonary function test report; A coordinate extraction module, configured to extract coordinate data for each flow rate capacity ring in the graphic data; a segment division module, configured to divide the coordinate data of the expiratory phase of each flow rate volume loop into different segments based on the coordinate data of the expiratory phase of each flow rate volume loop and according to the change characteristics and time characteristics of the flow rate; an abnormality determination module, configured to determine whether each of the flow rate and capacity loops has an abnormal exhalation condition based on the different segments of the coordinate data and whether the coordinate data form a preset shape feature, wherein different preset shape features correspond to different abnormal exhalation conditions, wherein the abnormal exhalation condition of insufficient initial exhalation effort is detected by combining persistent coherent topological features and curve geometric features; a consistency judgment module, configured to judge whether the examination parameter data of multiple pulmonary function tests in the PDF file meet result consistency based on trend differences of the examination parameter data; The segment division module is configured to determine a maximum flow rate value based on the coordinate data of the expiratory phase of each flow rate capacity loop; determine a key time point according to the expiratory duration required for a pulmonary function test, and determine the flow rate value corresponding to the key time point, wherein the key time point is the last time point during the expiratory duration at which the flow rate continues to increase; determine the minimum value between the maximum flow rate value and the flow rate value corresponding to the key time point as a turning point; and determine the coordinate data before the turning point as a first segment and the coordinate data after the turning point as a second segment based on the generation time sequence of the coordinate data of the expiratory phase; The abnormality determination module is configured to convert the coordinate data within the second segment from time domain data to frequency domain data, determine high-frequency components in the frequency domain data, construct a polar coordinate system, map the polar angle in the polar coordinate system to vital capacity, map the polar diameter in the polar coordinate system to flow velocity, input the vital capacity and flow velocity values ​​in the coordinate data corresponding to the high-frequency components into the polar coordinate system, and generate a polar coordinate curve by adjusting the time window length and scaling parameters; if the polar diameter of the polar coordinate curve has a first preset shape feature of decreasing and then increasing, then determine that the flow rate capacity loop has the exhalation abnormality of coughing; The abnormality judgment module is used to calculate the gradient of each coordinate data within the coordinate data in the first section, and the gradient of each coordinate data forms a gradient vector; construct a gradient autocorrelation matrix based on the gradient vector; calculate the eigenvector corresponding to the minimum eigenvalue based on the gradient autocorrelation matrix; determine the point where the gradient direction suddenly changes as a corner point according to the eigenvector corresponding to the minimum eigenvalue, and determine at least one corner point; form a line between the coordinate data corresponding to the at least one corner point and the coordinate data before and after it; if the line has a first preset shape feature of decreasing and then rising, it is determined that the flow rate capacity loop has the exhalation abnormality of hesitating at the start of exhalation.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the quality control method for the pulmonary function test report according to any one of claims 1 to 4 is implemented.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program for executing the quality control method for a pulmonary function test report according to any one of claims 1 to 4.

Images