Lung function specific disease interpretation device and computer equipment

Abstract

The invention discloses a lung function specific disease interpretation device and computer equipment, and relates to the technical field of computers.The device comprises a first obtaining module used for obtaining a user lung function examination report, segmenting information of the lung function examination report, obtaining a flow velocity capacity ring image and sending the flow velocity capacity ring image to a second obtaining module; the flow velocity capacity loop image comprises a plurality of flow velocity capacity loop curves; the second acquisition module is used for acquiring coordinate data of each flow velocity capacity loop curve; and the judgment module is used for analyzing curve characteristics of each flow velocity capacity loop curve according to the coordinate data, the curve characteristics comprise platform type characteristics and double-butterfly type characteristics, and whether the user has lung function specific diseases or not is judged based on the curve characteristics. By performing computer identification and interpretation on the examination report, the interpretation efficiency and the interpretation accuracy of the lung function specific disease are improved, and the diagnosis cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of medical diagnosis, and particularly relates to a device for interpreting pulmonary function specific diseases and a computer device. Background Art

[0002] Currently, in pulmonary function examinations, the interpretation of ventilation function examinations is based on the interpretation of pulmonary function parameter data. However, there are many specific diseases caused by pulmonary tumor diseases in pulmonary function examinations, which cannot be diagnosed through the traditional parameter data interpretation criteria. Only through the clinical experience of the operating doctor can the shape of the pulmonary function flow-volume loop be diagnosed. These experiences are not available to doctors in primary hospitals, and misdiagnosis and missed diagnosis are likely to occur. Therefore, through the method of computer machine learning, the computer can automatically diagnose the report graph, which can timely remind medical staff and prevent misdiagnosis and missed diagnosis. Summary of the Invention

[0003] In view of this, the present invention provides a device for interpreting pulmonary function specific diseases to solve the technical problem that doctors with insufficient experience are prone to misdiagnosis and missed diagnosis of pulmonary function specific diseases in the prior art. The device includes: A first acquisition module, configured to acquire a user's pulmonary function examination report, segment the information of the pulmonary function examination report, and obtain a flow-volume loop image, where the flow-volume loop image includes a plurality of flow-volume loop curves; A second acquisition module, configured to respectively obtain the coordinate data of each flow-volume loop curve; A judgment module, configured to respectively analyze the curve characteristics of each flow-volume loop curve according to the coordinate data for each flow-volume loop curve, where the curve characteristics include a platform type characteristic and a double butterfly type characteristic, and judge whether the user has a pulmonary function specific disease based on the curve characteristics.

[0004] Further, the first acquisition module is further configured to: segment the information of the pulmonary function examination report based on the PDFMiner parsing software to obtain a flow-volume loop image.

[0005] Further, the second acquisition module is further configured to: identify the text labels in the flow-volume loop image; locate the coordinate axes according to the text labels, obtain the coordinate origin and the coordinate scale values; obtain the pixel coordinates corresponding to the maximum value in the coordinate scale values; calculate the conversion ratio from the pixel coordinates to the physical quantity coordinates according to the pixel coordinates, the coordinate origin, and the maximum value in the coordinate scale values; and respectively obtain the physical quantity coordinates of each point on each flow-volume loop curve based on the conversion ratio.

[0006] Further, the determination module is further configured to: set a proportional threshold for flow rate, a proportional threshold for volume, and a curve slope threshold; obtain a first maximum flow rate value and a first maximum volume value of the flow rate-volume loop curve according to the coordinate data; obtain a flow rate threshold based on the first maximum flow rate value and the proportional threshold for flow rate; obtain a volume threshold based on the first maximum volume value and the proportional threshold for volume; start from a preset number of coordinate points after the coordinate point corresponding to the first maximum flow rate value, traverse the coordinate data until the coordinate point corresponding to the volume threshold is reached; if the flow rate value of a coordinate point is greater than or equal to the flow rate threshold and when the slope difference between the current coordinate point and the next coordinate point is less than or equal to the curve slope threshold, then count the platform-type feature; when the count of the platform-type feature is greater than a preset value, then the flow rate-volume loop curve has a platform-type feature.

[0007] Furthermore, the proportional threshold for flow rate is set to 0.5, the proportional threshold for volume is set to 0.5, and the curve slope threshold is set to 0.005.

[0008] Furthermore, when the count of the platform-type feature is greater than 4, then the flow rate-volume loop curve has a platform-type feature.

[0009] Furthermore, the determination module is further configured to: Traverse each flow rate-volume loop curve, and perform absolute value processing on the coordinate data; Process each flow rate-volume loop curve with an even index, and obtain a second maximum flow rate value and a second maximum volume value according to the coordinate data; Obtain a straight line between the first coordinate point corresponding to the second maximum flow rate value and the second coordinate point corresponding to the second maximum volume value; In the direction of the flow rate coordinate axis, calculate the total number of coordinate points on the flow rate-volume loop curve that are lower than the straight line; If the ratio of the total number of coordinate points lower than the straight line to the total number of coordinate points on the flow rate-volume loop curve exceeds 0.5, then this flow rate-volume loop curve with an even index has a part of the double butterfly-type feature; Process each flow rate-volume loop curve with an odd index, and obtain a third maximum flow rate value and a minimum volume value according to the coordinate data; Obtain a first area enclosed by the flow rate-volume loop curve and the volume coordinate axis between the third coordinate point corresponding to the third maximum flow rate value and the fourth coordinate point corresponding to the minimum volume value; Obtain a second area of a rectangle with the connection line between the third coordinate point and the fourth coordinate point as the diagonal; If the ratio of the difference between twice the first area and the second area to the second area is less than 0.25, then a part of the double butterfly-shaped feature exists in the flow volume loop curve of the odd index; If a part of the double butterfly-shaped feature exists in both the flow volume loop curve of the even index and the flow volume loop curve of the odd index, then the double butterfly-shaped feature exists in the flow volume loop curve.

[0010] Furthermore, the determination module is further configured to: When a plateau-shaped feature exists in the inhalation term of the flow volume loop curve, output a warning message for a specific pulmonary function disease of the user with variable extrathoracic airway obstruction; When a plateau-shaped feature exists in the exhalation term of the flow volume loop curve, output a warning message for a specific pulmonary function disease of the user with variable intrathoracic airway obstruction; When plateau-shaped features exist in both the inhalation term and the exhalation term of the flow volume loop curve, output a warning message for a specific pulmonary function disease of the user with fixed airway obstruction; When the double butterfly-shaped feature exists in the flow volume loop curve, output a warning message for a specific pulmonary function disease of the user with incomplete obstruction of the unilateral main bronchus.

[0011] Further, the device further includes: a push module, configured to push a disease risk warning message to the user if there is a specific pulmonary function disease.

[0012] The present invention also provides a computer device, including any of the above specific pulmonary function disease interpretation devices, to solve the technical problem that doctors with insufficient experience in the prior art are prone to missed diagnosis and misdiagnosis of specific pulmonary function diseases.

[0013] Compared with the prior art, the beneficial effects that can be achieved by at least one of the above technical solutions adopted in the embodiments of this specification at least include: a first acquisition module, configured to acquire a user's pulmonary function examination report, segment the information of the pulmonary function examination report to obtain a flow-volume loop image, where the flow-volume loop image includes a plurality of flow-volume loop curves; a second acquisition module, configured to respectively obtain the coordinate data of each flow-volume loop curve; a judgment module, configured to, for each flow-volume loop curve, respectively analyze the curve characteristics of the flow-volume loop curve according to the coordinate data, where the curve characteristics include a plateau type characteristic and a double butterfly type characteristic, and determine whether the user has a pulmonary function specific disease based on the curve characteristics. Through the method of computer machine learning, this application enables the computer to automatically diagnose the report graph, remind medical staff, and prevent missed diagnosis and misdiagnosis; this method can quickly and automatically analyze a large amount of pulmonary function data, improving the work efficiency of doctors; by the computer identifying the complex relationship between pulmonary function specific diseases and pulmonary function data and making an interpretation, the accuracy of disease interpretation is improved; at the same time, this interpretation device can help doctors with early screening and auxiliary diagnosis, reducing the medical costs of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0015] Figure 1 is a structural block diagram of a device for interpreting pulmonary function specific diseases provided by an embodiment of the present invention; Figure 2 is a schematic diagram of flow-volume loop curves corresponding to different types of pulmonary function specific diseases provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] The embodiments of the present application will be described in detail below with reference to the drawings.

[0017] The following describes the implementation manners of the present application through specific specific examples. Those skilled in the art can easily understand other advantages and effects of the present application from the content 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 implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts belong to the scope protected by the present application.

[0018] In an embodiment of the present invention, a lung function specific disease interpretation device is provided. As Figure 1 shown, it includes: a first acquisition module 101, a second acquisition module 102, and a judgment module 103. The following describes this structure.

[0019] The first acquisition module 101 is configured to acquire a user's lung function examination report, segment the information in the lung function examination report, and obtain a flow-volume loop image, where the flow-volume loop image includes a plurality of flow-volume loop curves; The second acquisition module 102 is configured to respectively obtain the coordinate data of each flow-volume loop curve; The judgment module 103 is configured to, for each flow-volume loop curve, respectively analyze the curve characteristics of the flow-volume loop curve according to the coordinate data, and judge whether the user has a lung function specific disease based on the curve characteristics.

[0020] In specific implementation, for the PDF file of the user's lung function examination report, it generally includes user basic information, lung function parameter data, and lung function image data. For specific diseases caused by lung tumor diseases, they cannot be diagnosed through the diagnostic criteria of lung function parameter data. The currently commonly used method is to diagnose by manually analyzing the shape of the flow-volume loop in the lung function image data. However, the manual diagnosis method is prone to the risks of missed diagnosis and misdiagnosis. In this embodiment, machine learning is used to assist in the diagnosis of lung function specific diseases, enabling the computer model to automatically diagnose the report graphics, which can quickly and automatically analyze a large amount of lung function data, improving the doctor's work efficiency; using the computer model to learn the complex relationship between diseases and lung function data, improving the accuracy of the interpretation of lung function specific diseases; at the same time, it can help doctors with early screening and auxiliary diagnosis, reducing the medical costs of patients.

[0021] In specific implementation, for the shape of the flow-volume loop curve of lung function specific diseases, refer to Figure 2, when there is a plateau feature in the inhalation phase of the flow-volume loop curve, the user has a specific pulmonary function disease of variable extrathoracic airway obstruction; when there is a plateau feature in the exhalation phase of the flow-volume loop curve, the user has a specific pulmonary function disease of variable intrathoracic airway obstruction; when there are plateau features in both the inhalation and exhalation phases of the flow-volume loop curve, the user has a specific pulmonary function disease of fixed airway obstruction; when the flow-volume loop curve has a double butterfly feature, the user has a specific pulmonary function disease of incomplete obstruction of the unilateral main bronchus. Therefore, the curve features of the curve can be recognized by a computer to help doctors identify whether the user has a risk of specific pulmonary function diseases.

[0022] In one embodiment, the first acquisition module 101 is further configured to: Segment the information in the pulmonary function examination report based on the PDFMiner parsing software to obtain a flow-volume loop image.

[0023] Specifically, PDFMiner is a Python library dedicated to parsing and extracting the content of PDF documents. It extracts text, fonts, images, and other metadata information by analyzing the internal structure of the PDF file of the pulmonary function examination report. In this embodiment, it mainly extracts the basic information of the user, the pulmonary function parameter data, and the pulmonary function image data.

[0024] In one embodiment, the second acquisition module 102 is further configured to: Identify the text labels in the flow-volume loop image; Locate the coordinate axes according to the text labels, and obtain the coordinate origin and the coordinate scale values; Obtain the pixel coordinates corresponding to the maximum value in the coordinate scale values; Calculate the conversion ratio from pixel coordinates to physical quantity coordinates according to the pixel coordinates, the coordinate origin, and the maximum value in the coordinate scale values; Based on the conversion ratio, obtain the physical quantity coordinates of each point on each flow-volume loop curve respectively.

[0025] In specific implementation, OCR recognition technology is used for recognition, including character recognition, Chinese character recognition, graphic recognition, etc., to parse out the axis labels in the flow-volume loop image, and locate the axes according to the axis labels. Specifically, for example, first locate the Y-axis label (flow velocity axis label), then extract the numerical text as the Y-axis coordinate scale value, obtain the Y-axis pixel coordinate corresponding to the maximum value in the Y-axis coordinate scale value, and then combine the Y-axis pixel coordinate of the coordinate origin to obtain the conversion ratio from the Y-axis pixel coordinate to the physical quantity coordinate. Similarly, the conversion ratio from the X-axis (volume axis) pixel coordinate to the physical quantity coordinate can be obtained. Based on this conversion ratio, the physical quantity coordinates of each point on each flow-volume loop curve can be obtained respectively.

[0026] In this embodiment, based on computer machine vision technology, the automatic recognition of lung function graphics in PDF files is realized, and the coordinate data of the flow-volume loop curve that can be used for subsequent analysis is generated.

[0027] In specific implementation, the obtaining of the coordinate data of each flow-volume loop curve respectively further includes: For each flow-volume loop curve after coordinate conversion, the data density of the curve is increased by using linear interpolation.

[0028] In one embodiment, the determination module 103 is further configured to: Set the ratio threshold of the flow velocity, the ratio threshold of the volume, and the curve slope threshold; According to the coordinate data, obtain the first maximum flow velocity value and the first maximum volume value of the flow-volume loop curve; Based on the first maximum flow velocity value and the ratio threshold of the flow velocity, obtain the flow velocity threshold; Based on the first maximum volume value and the ratio threshold of the volume, obtain the volume threshold; Starting from a preset number of coordinate points after the coordinate point corresponding to the first maximum flow velocity value, traverse the coordinate data until the coordinate point corresponding to the volume threshold is reached; For the flow velocity value of the coordinate point greater than or equal to the flow velocity threshold, and when the slope difference between the current coordinate point and the next coordinate point is less than or equal to the curve slope threshold, the counting of the platform-type feature is performed; When the count of the platform-type feature is greater than the preset value, the flow-volume loop curve has a platform-type feature.

[0029] In one embodiment, the ratio threshold of the flow velocity is set to 0.5, the ratio threshold of the volume is set to 0.5, and the curve slope threshold is set to 0.005.

[0030] In one embodiment, when the count of the platform - type feature is greater than 4, there is a platform - type feature in the flow rate - capacity loop curve.

[0031] In specific implementation, the identification of the platform - type feature can be achieved through the following process, which specifically includes the following content: if P_Sign: # Platform - type feature detection PEF_Ratio,FVC_Ratio = 0.5, 0.5 # Set the ratio thresholds of PEF and FVC to 0.5 respectively Slope_Thresh = 0.005 # Set the slope threshold of 0.005 to judge the platform Positive_Sign = False # Mark whether there is a platform - type feature for c_c in range(0, len(X_Store), 2): # Traverse each curve, with a step size of 2 curve_x, curve_y = X_Store[c_c], Y_Store[c_c] # Obtain the x and y data of the current curve FVC_Max, PEF_Max, PEF_Max_Loc = np.max(curve_x), p.max(curve_y), np.argmax(curve_y) # Find the maximum FVC, PEF and their positions PEF_Thresh, FVC_Thresh_Loc = PEF_Max * PEF_Ratio, nt(len(curve_x) * FVC_Ratio) # Calculate the PEF threshold and the FVC threshold position slope_record, platform_count, platform_max, pef_min, pef_max, fvc_min, fvc_max = [], 0, 0, 0, 0, 0, 0 # Initialize variables start_loc, end_loc = 0, 0 # Initialize the start and end positions for y_c in range(PEF_Max_Loc + 3, FVC_Thresh_Loc): # Traverse from the third point after the PEF maximum position to the FVC threshold position if curve_y[y_c] < PEF_Thresh: # If the y - value of the current point is lower than the PEF threshold, break the loop break slope_record.append(curve_y[y_c] - curve_y[y_c + 1]) # Record the slope difference between the current point and the next point if curve_y[y_c] - curve_y[y_c + 1] > Slope_Thresh: # If the slope difference is greater than the threshold if platform_count > platform_max: # Update the information of the longest platform platform_max = platform_count start_loc = start_loc_temp end_loc = y_c platform_count = 0 # Reset the platform counter else: if platform_count == 0: # If a new platform starts, record the starting position start_loc_temp = y_c platform_count += 1 # Increase the platform counter if platform_max >= 4: # If there is a platform with a length greater than or equal to 4 print('Platform Positive') # Output that the platform feature exists Positive_Sign = True # Set the flag pef_max, pef_min, fvc_max, fvc_min = curve_y[start_loc], curve_y[end_loc], 100.0 * end_loc / len(curve_x), 100.0 * start_loc / len(curve_x) # Calculate the maximum and minimum values of PEF and FVC print('PEF_High: {}\t PEF_Low:{}'.format(pef_max, pef_min)) # Output the high and low values of PEF print('FVC_Max: {}\t FVC_Min:{}'.format(fvc_max, fvc_min)) # Output the maximum and minimum percentages of FVC break # End the loop if not Positive_Sign: print('Platform Negative') # Output that the platform feature does not exist The above process implements a platform feature detection algorithm for analyzing flow-volume loop curve data. Specifically, it identifies whether there is a platform feature by traversing the curve data and calculating the slope difference. The following are the main steps: (1) Initialize parameters and variables: Set the ratio thresholds for PEF (peak expiratory flow) and FVC (forced vital capacity) (PEF_Ratio and FVC_Ratio); Set the slope threshold (Slope_Thresh) for determining the platform region; Initialize the flag variable Positive_Sign to record whether there is a platform feature; (2) Traverse each curve: Traverse the curve data stored in X_Store and Y_Store with a step size of 2; For each curve, obtain its x and y data, and find the maximum FVC, PEF, and their positions; Calculate the PEF threshold (PEF_Thresh) and the FVC threshold position (FVC_Thresh_Loc); (3) Find the platform region: Start traversing from the third point after the PEF maximum position to the FVC threshold position; Calculate the slope difference between the current point and the next point for each point and record it in the slope_record list; If the slope difference is greater than the set slope threshold, it is considered that the current point is not in the platform region, and reset the platform counter; otherwise, increment the platform counter; Update the information of the longest platform (starting position start_loc_temp and ending position end_loc), and set the flag Positive_Sign when the platform length exceeds 4 points; (4) Output the result: If a platform region that meets the conditions is found, output "Platform Positive" and print the maximum and minimum values of the relevant PEF and FVC; If a platform region that meets the conditions is not found, output "Platform Negative".

[0032] In this embodiment, by calculating the slope difference to determine the platform area, the platform-type features can be more accurately identified, avoiding misjudgments that may be caused by simple threshold judgments; users are allowed to adjust the ratio threshold and slope threshold of PEF and FVC according to actual needs, making the algorithm highly flexible and adaptable; while identifying the platform, the maximum and minimum values of PEF and FVC and their relative positions are retained, which helps to further analyze and interpret the curve data; clear output information is provided, including whether there are platform-type features, the maximum and minimum values of PEF and FVC, etc., facilitating users to understand and use.

[0033] In one embodiment, the determination module 103 is further configured to: Traverse each flow rate-capacity loop curve and perform absolute value processing on the coordinate data; Process each flow rate-capacity loop curve with an even index, and obtain the second maximum flow rate value and the second maximum capacity value according to the coordinate data; According to the first coordinate point corresponding to the second maximum flow rate value and the second coordinate point corresponding to the second maximum capacity value, obtain the straight line between the first coordinate point and the second coordinate point; In the direction of the flow rate coordinate axis, calculate the total number of coordinate points on the flow rate-capacity loop curve that are lower than the straight line; If the ratio of the total number of coordinate points lower than the straight line to the total number of coordinate points on the flow rate-capacity loop curve exceeds 0.5, then a part of the double butterfly-type feature exists in the flow rate-capacity loop curve with this even index; Process each flow rate-capacity loop curve with an odd index, and obtain the third maximum flow rate value and the minimum capacity value according to the coordinate data; According to the third coordinate point corresponding to the third maximum flow rate value and the fourth coordinate point corresponding to the minimum capacity value, obtain the first area enclosed by the flow rate-capacity loop curve and the capacity coordinate axis between the third coordinate point and the fourth coordinate point; Obtain the second area of the rectangle with the connection line between the third coordinate point and the fourth coordinate point as the diagonal; If the ratio of the difference between twice the first area and the second area to the second area is less than 0.25, then a part of the double butterfly-type feature exists in the flow rate-capacity loop curve with this odd index; If a part of the double butterfly-type feature exists in both the flow rate-capacity loop curve with an even index and the flow rate-capacity loop curve with an odd index, then the flow rate-capacity loop curve has a double butterfly-type feature.

[0034] Specifically, when implemented, the identification of the double butterfly-type feature can be achieved through the following process, which specifically includes the following content: if D_Sign: # Double Butterfly Feature Detection butterfly_record, butterfly_sign = np.zeros((2, 1)), False # Initialize the double butterfly feature record and flag for l_c in range(0, len(X_Store)): # Traverse each curve curve_x, curve_y = np.abs(X_Store[l_c]), np.abs(Y_Store[l_c]) # Get the absolute value of x and y data of the current curve if l_c % 2 == 0: # Process the curve with an even index butterfly_record = np.zeros((2, 1)) # Reset the double butterfly feature record start_loc, end_loc = np.argmax(curve_y), np.argmax(curve_x) # Find the positions of the maximum y value and maximum x value x1, y1, x2, y2 = curve_x[start_loc], curve_y[start_loc], curve_x[end_loc], curve_y[end_loc] # Get the coordinates of the corresponding points k = (y2 - y1) / (x2 - x1) # Calculate the slope of the straight line b = y1 - k * x1 # Calculate the y-intercept of the straight line lower_pts_count = 0 # Initialize the count of points below the line for p_c in range(start_loc, end_loc): # Traverse the points from the maximum y value to the maximum x value y_line = k * curve_x[p_c] + b # Calculate the corresponding y value on the line y_real = curve_y[p_c] # Actual y value if y_real < y_line: # If the actual y value is below the line lower_pts_count += 1 # Increase the count lower_ratio = lower_pts_count / (end_loc - start_loc) # Calculate the ratio of points below the line if lower_ratio > 0.5: # If the ratio of points below the line exceeds half butterfly_record[int(l_c % 2)] = 1 # Record the double butterfly feature else: # Process the curve with an odd index start_loc, end_loc = np.argmax(curve_y), np.argmin(curve_x) # Find the positions of the maximum y value and the minimum x value pt_x, pt_y = curve_x[start_loc:end_loc], curve_y[start_loc: end_loc] # Obtain the x and y data for the corresponding interval auc_real = auc(pt_x, pt_y) # Calculate the area under the interval auc_square = abs(curve_x[start_loc] - curve_x[end_loc]) * abs(curve_y[start_loc] - curve_y[end_loc]) # Calculate the rectangle area if (2 * auc_real - auc_square) / auc_square < 0.25: # If the condition is met butterfly_record[int(l_c % 2)] = 1 # Record the double butterfly feature if np.sum(butterfly_record) == 2: # If both curves meet the double butterfly feature condition butterfly_sign = True # Set the flag if butterfly_sign: print('Double Butterfly Positive.') # Output that the double butterfly feature exists else: print('Double Butterfly Negative.') # Output that the double butterfly feature does not exist The above process implements a Double Butterfly feature detection algorithm for analyzing flow volume loop curve data. Specifically, it identifies the presence of the Double Butterfly feature by calculating the slope and area. The following are the main steps: (1) Initialize parameters and variables: butterfly_record: A 2x1 array used to record whether the Double Butterfly feature is found; butterfly_sign: A boolean variable used to mark the presence of the Double Butterfly feature; (2) Traverse each curve: Traverse the curve data stored in X_Store and Y_Store; For each curve, obtain its absolute value x and y data and perform different processing according to the index (even or odd); (3) Process curves with even indices: Find the positions of the maximum y-value point and the maximum x-value point, and obtain the corresponding coordinates (x1, y1) and (x2, y2); Calculate the slope k and intercept b of the line connecting these two points; Traverse all points between the maximum y-value point and the maximum x-value point, and count the number of points where the actual y-value is lower than the corresponding y-value on the line; Calculate the ratio of these points. If the ratio exceeds 50%, record the Double Butterfly feature; (3) Process curves with odd indices: Find the positions of the maximum y-value point and the minimum x-value point, and obtain the x and y data for the corresponding interval; Calculate the area under the curve (auc_real) and the rectangular area (auc_square); If (2 * auc_real - auc_square) / auc_square is less than 0.25, record the Double Butterfly feature; (4) Determine the presence of the Double Butterfly feature: If both curves meet the Double Butterfly feature conditions, set the flag butterfly_sign to True; Output the corresponding information according to the value of butterfly_sign.

[0035] In this embodiment, for curves with even indexes, the feature is determined by counting the proportion of points below the straight line; for curves with odd indexes, the feature is determined by comparing the area under the interval with the area of ​​the rectangle, which can more accurately identify this complex morphological feature. While identifying the double butterfly feature, the layout information of the original curve is retained, which is helpful for further analysis and interpretation of the data; clear output information is provided, including whether the double butterfly feature exists, which is easy for users to understand and use.

[0036] In one embodiment, the determining module 103 is further configured to: When the inspiratory item of the flow rate volume loop curve has a plateau feature, outputting a warning message that the user has a pulmonary function-specific disease of variable extrathoracic airway obstruction; When the exhalation item of the flow rate volume loop curve has a plateau feature, outputting a warning message that the user has a pulmonary function-specific disease of variable intrathoracic airway obstruction; When both the inspiratory item and the expiratory item of the flow rate volume loop curve have plateau characteristics, outputting warning information that the user has a lung function-specific disease with fixed airway obstruction; When the flow rate volume loop curve has a double butterfly-shaped feature, a warning message is outputted indicating that the user has a lung function-specific disease of incomplete unilateral main bronchial obstruction.

[0037] During specific implementation, the above interpretation results are pushed to the user. If they are relevant interpretation results, an information reminder is pushed to the user, indicating that the user's report has the risk of lung function-specific diseases and further CT examination is required to eliminate the risk.

[0038] In one embodiment, the apparatus further comprises: The push module is used to push disease risk warning information to users if lung function-specific diseases exist.

[0039] The lung function-specific disease interpretation device of the present application has the following characteristics: Improve interpretation efficiency: Using computers to interpret lung function-specific diseases can quickly and automatically analyze large amounts of lung function data, improving doctors' work efficiency; Improved interpretation accuracy: Computer models can learn the complex relationship between disease and lung function data, improving the accuracy of disease interpretation; Reduce diagnostic costs: Computer models can help doctors conduct early screening and auxiliary diagnosis, reducing patients' medical costs.

[0040] In another embodiment, a computer device is provided, comprising any of the above-mentioned pulmonary function-specific disease interpretation devices.

[0041] Specifically, the computer device may be a computer terminal, a server, or a similar computing device.

[0042] The embodiments of the present invention achieve the following technical effects: a first acquisition module for acquiring a user's pulmonary function examination report, segmenting the information in the pulmonary function examination report to obtain a flow-volume loop image, where the flow-volume loop image includes multiple flow-volume loop curves; a second acquisition module for respectively obtaining the coordinate data of each flow-volume loop curve; a judgment module for, for each flow-volume loop curve, respectively analyzing the curve characteristics of the flow-volume loop curve according to the coordinate data, where the curve characteristics include a plateau type characteristic and a double butterfly type characteristic, and determining whether the user has a pulmonary function specific disease based on the curve characteristics. This application enables the computer to automatically diagnose the report graphics through computer machine learning, reminding medical staff to prevent missed diagnoses and misdiagnoses; this method can quickly and automatically analyze a large amount of pulmonary function data, improving the work efficiency of doctors; by the computer identifying the complex relationship between pulmonary function specific diseases and pulmonary function data and making an interpretation, the accuracy of disease interpretation is improved; at the same time, this interpretation device can help doctors with early screening and auxiliary diagnosis, reducing the medical costs of patients.

[0043] Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the embodiments of the present invention can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. Optionally, they can be implemented by program codes executable by the computing device, so that they can be stored in a storage device and executed by the computing device. And in some cases, the steps shown or described can be executed in a different order than here, or they can be separately made into individual integrated circuit modules, or multiple of them can be made into a single integrated circuit module to implement. Thus, the embodiments of the present invention are not limited to any specific combination of hardware and software.

[0044] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the embodiments of the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A lung function-specific disease interpretation device, characterized in that, Including: A first acquisition module, configured to acquire a user's pulmonary function examination report, segment the information in the pulmonary function examination report to obtain a flow-volume loop image, and the flow-volume loop image includes a plurality of flow-volume loop curves; A second acquisition module, configured to respectively obtain the coordinate data of each flow-volume loop curve; A judgment module, configured to, for each flow-volume loop curve, respectively analyze the curve characteristics of the flow-volume loop curve according to the coordinate data, where the curve characteristics include a plateau type characteristic and a double butterfly type characteristic, and determine whether the user has a pulmonary function specific disease based on the curve characteristics.

2. The lung function specific disease interpretation device according to claim 1, wherein The first acquisition module is further configured to: Segment the information in the pulmonary function examination report based on the PDFMiner parsing software to obtain a flow-volume loop image.

3. The pulmonary function specific disease interpretation device according to claim 1, wherein The second acquisition module is further configured to: Identify the text labels in the flow-volume loop image; Locate the coordinate axes according to the text labels, and obtain the coordinate origin and the coordinate scale values; Obtain the pixel coordinates corresponding to the maximum value in the coordinate scale values; Calculate the conversion ratio from pixel coordinates to physical quantity coordinates according to the pixel coordinates, the coordinate origin, and the maximum value in the coordinate scale values; Based on the conversion ratio, respectively obtain the physical quantity coordinates of each point on each flow-volume loop curve.

4. The pulmonary function specific disease interpretation device according to claim 1, wherein The judgment module is further configured to: Set a ratio threshold for flow rate, a ratio threshold for volume, and a curve slope threshold; According to the coordinate data, obtain the first maximum flow rate value and the first maximum volume value of the flow-volume loop curve; Based on the first maximum flow rate value and the ratio threshold for flow rate, obtain a flow rate threshold; Based on the first maximum volume value and the ratio threshold for volume, obtain a volume threshold; Starting from a preset number of coordinate points after the coordinate point corresponding to the first maximum flow rate value, traverse the coordinate data until the coordinate point corresponding to the volume threshold is reached; For the flow rate value of the coordinate point being greater than or equal to the flow rate threshold, and when the slope difference between the current coordinate point and the next coordinate point is less than or equal to the curve slope threshold, count the plateau type characteristic; When the count of the plateau type characteristic is greater than a preset value, the flow-volume loop curve has a plateau type characteristic.

5. The lung function specific disease interpretation device according to claim 4, wherein The ratio threshold for flow rate is set to 0.5, the ratio threshold for volume is set to 0.5, and the curve slope threshold is set to 0.

005.

6. The pulmonary function specific disease interpretation device according to claim 4, wherein, When the count of the plateau type characteristic is greater than 4, the flow-volume loop curve has a plateau type characteristic.

7. The pulmonary function specific disease interpretation device according to claim 4, characterized in that, The judgment module is further configured to: Traverse each flow-volume loop curve and perform absolute value processing on the coordinate data; Process each flow-volume loop curve with an even index, and obtain a second maximum flow rate value and a second maximum volume value according to the coordinate data; Obtain the straight line between the first coordinate point corresponding to the second maximum flow rate value and the second coordinate point corresponding to the second maximum volume value; In the direction of the flow rate coordinate axis, calculate the total number of coordinate points on the flow-volume loop curve that are lower than the straight line. If the ratio of the total number of coordinate points below the straight line to the total number of coordinate points on the flow volume loop curve exceeds 0.5, then a part of the double butterfly feature exists in the flow volume loop curve with an even index; Process each flow volume loop curve with an odd index, and obtain the third maximum flow rate value and the minimum volume value according to the coordinate data; According to the third coordinate point corresponding to the third maximum flow rate value and the fourth coordinate point corresponding to the minimum volume value, obtain the first area enclosed by the flow volume loop curve between the third coordinate point and the fourth coordinate point and the volume coordinate axis; Obtain the second area of the rectangle with the line connecting the third coordinate point and the fourth coordinate point as the diagonal; If the ratio of the difference between twice the first area and the second area to the second area is less than 0.25, then a part of the double butterfly feature exists in the flow volume loop curve with an odd index; If a part of the double butterfly feature exists in both the flow volume loop curve with an even index and the flow volume loop curve with an odd index, then the double butterfly feature exists in the flow volume loop curve.

8. The lung function specific disease interpretation device according to claim 7, wherein, The determination module is further configured to: When a plateau feature exists in the inhalation phase of the flow volume loop curve, output a warning message for a specific pulmonary function disease of the user with variable extrathoracic airway obstruction; When a plateau feature exists in the exhalation phase of the flow volume loop curve, output a warning message for a specific pulmonary function disease of the user with variable intrathoracic airway obstruction; When plateau features exist simultaneously in the inhalation phase and the exhalation phase of the flow volume loop curve, output a warning message for a specific pulmonary function disease of the user with fixed airway obstruction; When the double butterfly feature exists in the flow volume loop curve, output a warning message for a specific pulmonary function disease of the user with incomplete obstruction of the unilateral main bronchus.

9. The lung function specific disease interpretation device according to claim 1, wherein, The device further includes: A push module, configured to push a disease risk warning message to the user if a specific pulmonary function disease exists.

10. A computer device, characterized in that, Including the specific pulmonary function disease interpretation device according to any one of claims 1 to 9.

Images