Calculation method of diaphragm surface smoothness based on Lipschitz index

Abstract

The present application proposes a method for calculating the smoothness of the diaphragm surface based on the Lipschitz index, which includes: extracting the lung contour; extracting the lung feature points; extracting the diaphragm surface contour line; calculating the diaphragm surface Lipschitz index; and taking the average of the Lipschitz indexes of each feature point calculated as the diaphragm surface smoothness. The present application uses the Lipschitz index to quantitatively measure the diaphragm surface smoothness, overcoming the deviation caused by the doctor's subjective experience; the larger the Lipschitz index, the less smooth the diaphragm surface, and the smaller the Lipschitz index, the smoother the diaphragm surface, providing an objective reference for clinical doctors to determine the degree of diaphragm surface smoothness.

Description

Technical Field

[0001] The present application relates to a method for calculating diaphragm surface smoothness. Background Art

[0002] The diaphragmatic surface is located below the lungs and above the top of the diaphragm. Due to the compression of the diaphragm, it is a crescent-shaped depression. Diaphragmatic surface smoothness refers to the smoothness of the diaphragmatic surface. When the diaphragmatic surface is not smooth, it may be related to pleural adhesion or pleural effusion. Patients with atelectasis may also have imaging manifestations such as the disappearance of the diaphragmatic surface. However, at present, clinicians' detection of diaphragmatic surface smoothness is judged by naked eye observation, and then a subjective description is given, indicating that the diaphragmatic surface is smooth or uneven.

[0003] Since clinicians make diagnoses through visual inspection, the accuracy of the results is often closely related to the doctors' clinical experience. The diagnostic levels of doctors in backward areas are also uneven, and the accuracy and stability of the results cannot be guaranteed. Summary of the invention

[0004] In view of the above problems, the present application aims to propose a method for calculating diaphragm surface smoothness based on the Lipschitz index.

[0005] The diaphragm surface smoothness calculation method based on the Lipschitz index of the present application includes:

[0006] Extract lung contours;

[0007] Extract lung feature points;

[0008] Extract the diaphragm contour line;

[0009] Calculate the Lipschitz index of the diaphragmatic surface;

[0010] The average value of the Lipschitz index of each feature point calculated was taken as the diaphragm surface smoothness.

[0011] Preferably, the SEDUCM method is used in the extraction of lung contours; in the SEDUCM method, the input DR image is decomposed by a guided filter and input into the base layer and detail layer of the SED, and then the SED model is trained to generate a boundary map for detecting lung boundaries; the thoracic cavity and spine centerlines are extracted from the boundary map and the input CXR image, and the CXR image is divided into the left thoracic cavity and the right thoracic cavity; then, the MWT algorithm and the UCM algorithm are used to generate candidate lung regions and contours, and finally the contours with the highest confidence are selected as the left and right lung contours.

[0012] Preferably, in extracting lung feature points, the highest point, the lowest point, and the left and right edge vertices of the left and right lungs are extracted based on the positional relationship; the lowest point is used as the costophrenic angle vertex; the coordinates of the center point of the edge line of the lung close to the spine are calculated based on the highest point and the lowest point, the one-third point between the costophrenic angle vertex and the center point is calculated, the lower edge line between the one-third point and the center point is extracted, the slope of the straight line formed by the costophrenic angle vertex and the points on the edge line is calculated, and the point with the smallest slope is taken as the costophrenic angle vertex.

[0013] Preferably, in extracting the diaphragm surface contour line, according to the coordinates of the apex of the costophrenic angle and the apex of the cardiophrenic angle, the lung edge line between the two points is extracted as the diaphragm surface contour line.

[0014] Preferably, in calculating the diaphragmatic surface Lipschitz index:

[0015] Choose an appropriate analytical wavelet;

[0016] For each feature point on the diaphragm contour line, find the local maximum point near the feature point whose position does not change with the measurement variable; use the exponential fitting to find the modulus of the local maximum point whose position does not change with the measurement variable, and further obtain its Lipschitz index.

[0017] Preferably, the analytical wavelet is a Mexican hat function, which is a second-order derivative of a Gaussian function.

[0018] This application uses the Lipschitz index to quantitatively measure the smoothness of the diaphragm surface, overcoming the deviation caused by the doctor's subjective experience; the larger the Lipschitz index, the less smooth the diaphragm surface, and the smaller the Lipschitz index, the smoother the diaphragm surface, providing an objective reference for clinicians to determine the smoothness of the diaphragm surface. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of the method for calculating diaphragm surface smoothness based on the Lipschitz index of the present application;

[0020] Figure 2 It is the flow chart of SEDUCM algorithm;

[0021] Figure 3 Schematic diagram of the locations of characteristic points of the lungs;

[0022] Figure 4 It is a schematic diagram of Mexican hat wavelet;

[0023] Figure 5 Schematic diagram of lung contour extraction. DETAILED DESCRIPTION

[0024] The method for calculating the diaphragm surface smoothness based on the Lipschitz index of the present application is described in detail below in conjunction with the accompanying drawings.

[0025] The method of the present application includes four steps. The first step is lung contour extraction, the second step is lung feature point extraction, the third step is diaphragm surface contour line extraction, and the fourth step is diaphragm surface smoothness calculation.

[0026] Step 1: Lung contour extraction

[0027] The lung contour extraction adopts the SEDUCM (structured edge detector-ultrametric contourmap) method. The input DR image is decomposed by the guided filter and input into the base layer and detail layer of SED. Then the SED model is trained to generate a boundary map for detecting the lung boundary. The thoracic cavity and spine centerline are extracted from the boundary map and the input CXR, and the CXR image is divided into the left thorax and the right thorax. Subsequently, the marker-controlled watershed transform (MWT) and ultrametric contour map (UCM) algorithm, namely mwt-ucm, are used to generate candidate lung regions and contours, and finally the contour with the highest confidence is selected as the left and right lung contours.

[0028] Step 2: Extract lung feature points

[0029] According to the positional relationship, the highest point, lowest point, and left and right edge vertices of the left and right lungs were extracted, and the lowest point was the apex of the costophrenic angle. The coordinates of the center point of the edge line of the lung close to the spine were calculated based on the highest and lowest points, and the one-third point between the apex of the costophrenic angle and the center point was calculated. The lower edge line between the one-third point and the center point was extracted, and the slope of the straight line formed by the apex of the costophrenic angle and each point on the edge line was calculated. The point with the smallest slope was taken and recorded as the apex of the costophrenic angle.

[0030] Step 3: Extract the diaphragm contour

[0031] According to the coordinates of the apex of the costophrenic angle and the apex of the cardiophrenic angle, the lung edge line between the two points is extracted, which is the diaphragmatic surface contour line.

[0032] Step 4: Calculate the diaphragmatic Lipschitz index

[0033] (1) Select a suitable analytical wavelet. The Mexican hat function, the second-order derivative of the Gaussian function, is used as the analytical wavelet. Its function expression is shown in equations 1 and 2, and its graph is shown in Figure 4 shown.

[0034]

[0035]

[0036] (2) Take a characteristic point x0 on the diaphragm contour line;

[0037] (3) Find the position near the feature point that does not change with the measurement variable a→0 + The local maximum point of the transformation, where a is the wavelet scaling variable;

[0038] (4) Use exponential fitting to find the modulus of the local extreme point whose position remains unchanged, according to Estimate the Lipschitz exponent α(x0), where b is the wavelet translation variable and C is a constant;

[0039] (5) Traverse all feature points on the diaphragm contour line and calculate the mean of the Lipschitz index of each feature point.

[0040] Next, combine Figure 5 This application is described.

[0041] Step 1: Lung contour extraction

[0042] Input lung frontal chest X-ray data (DICOM format), perform mean filter preprocessing, extract lung contour, and the contour map is as follows: Figure 5 shown.

[0043] Step 2: Extract lung feature points

[0044] (1) Extraction of the highest point of the lung: Extract the point with the largest ordinate in the contour lines of the left and right lungs;

[0045] (2) Extraction of the apex of the costophrenic angle: The costophrenic angle is located at the lowest point of the lung. The point with the smallest ordinate in the contour lines of the left and right lungs is extracted as the apex of the costophrenic angle.

[0046] (3) Extraction of the cardiodiaphragmatic angle vertex: Calculate the longitudinal midpoint of the left and right lung contours respectively, calculate the one-third point between the costophrenic angle vertex and the center point, extract the lower edge line between the one-third point and the center point, calculate the slope of the straight line formed by the costophrenic angle vertex and each point on the edge line, and take the maximum and minimum slopes for the left and right lungs respectively. The corresponding two points are the cardiodiaphragmatic points of the two lungs;

[0047] Step 3: Extraction of diaphragm contour

[0048] Extract the lower lung contour line between the apex of the costophrenic angle and the apex of the cardiophrenic angle, which is the diaphragmatic surface contour line;

[0049] Step 4: Calculate the Lipschitz index

[0050] Traverse each point on the diaphragm contour line and calculate the Lipschitz index.

[0051] The average value is the diaphragm surface smoothness index of this example.

[0052] Unless otherwise defined, all technical and / or scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which the invention relates. The materials, methods and examples mentioned in this application are illustrative only and not limiting.

[0053] Although the present invention has been described in conjunction with specific embodiments, within the spirit of the invention of this application, those skilled in the art may make appropriate substitutions, modifications and changes, and such substitutions, modifications and changes still fall within the scope of protection of this application.

Claims

1. A method for calculating diaphragm surface smoothness based on Lipschitz index, comprising: Extract lung contours; Extract lung feature points; Extract the diaphragm contour line; Calculate the Lipschitz index of the diaphragmatic surface; The average value of the Lipschitz index of each feature point calculated is taken as the diaphragm surface smoothness; In calculating the diaphragmatic Lipschitz index: Choose an appropriate analytical wavelet; For each feature point on the diaphragm contour line, find the local maximum point near the feature point whose position does not change with the measurement variable; use the exponential fitting to find the modulus of the local maximum point whose position does not change with the measurement variable, and further obtain its Lipschitz index.

2. The method for calculating diaphragm surface smoothness based on Lipschitz index according to claim 1, characterized in that: The SEDUCM method is used to extract lung contours. In the SEDUCM method, the input DR image is decomposed by a guided filter and input into the base layer and detail layer of SED. Then the SED model is trained to generate a boundary map for detecting lung boundaries. The thoracic cavity and spine centerlines are extracted from the boundary map and the input CXR image, and the CXR image is divided into the left thoracic cavity and the right thoracic cavity. Subsequently, the MWT algorithm and the UCM algorithm are used to generate candidate lung regions and contours, and finally the contours with the highest confidence are selected as the left and right lung contours.

3. The method for calculating diaphragm surface smoothness based on Lipschitz index according to claim 2, characterized in that: In extracting lung feature points, the highest point, lowest point, and left and right edge vertices of the left and right lungs are extracted according to the positional relationship; the lowest point is used as the costophrenic angle vertex; the coordinates of the center point of the edge line of the lung close to the spine are calculated based on the highest point and the lowest point, the one-third point between the costophrenic angle vertex and the center point is calculated, the lower edge line between the one-third point and the center point is extracted, the slope of the straight line formed by the costophrenic angle vertex and the points on the edge line is calculated, and the point with the smallest slope is taken as the costophrenic angle vertex.

4. The method for calculating diaphragm surface smoothness based on Lipschitz index according to claim 3, characterized in that: In extracting the diaphragm contour, according to the coordinates of the apex of the costophrenic angle and the apex of the cardiophrenic angle, the lung edge line between the two points is extracted as the diaphragm contour.

5. The method for calculating diaphragm surface smoothness based on Lipschitz index according to claim 1, characterized in that: The analytical wavelet is the Mexican hat function of the second-order derivative of the Gaussian function.

Images