Lung field image segmentation method, device and storage medium

Abstract

The present invention discloses a lung field image segmentation method, device, and storage medium. The method comprises: determining the gradient amplitude of a preprocessed chest image and integrating the gradient amplitude; determining a chest image based on the integration result, and segmenting the chest image into a left chest image and a right chest image based on a left lung segmentation point and a right lung segmentation point; performing costal margin segmentation, lung apex segmentation, and transverse and mediastinum segmentation on the left chest image and the right chest image, respectively, to obtain left and right costal margin segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinum segmentation images; determining lung apex region boundaries, transverse mediastinum region boundaries, and costal margin region boundaries based on the left and right costal margin segmentation images, the left and right lung apex segmentation images, and the left and right transverse and mediastinum segmentation images, and connecting the lung apex region boundaries, transverse mediastinum region boundaries, and costal margin region boundaries to obtain a lung field image. The present invention segments and connects the costal margin, lung apex, and transverse mediastinum regions of the chest, thereby improving segmentation accuracy.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a lung field image segmentation method, device and storage medium. Background Art

[0002] Currently, lung field segmentation technology only extracts the contour of the entire lung and is unable to perform fine segmentation analysis on the lung field contour, resulting in low accuracy of the segmented lung field image.

[0003] The above content is only used to assist in understanding the technical solution of the present invention and does not constitute an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of the present invention is to provide a lung field image segmentation method, device and storage medium, aiming to solve the technical problem of how to improve the accuracy of lung field image segmentation.

[0005] To achieve the above object, the present invention provides a lung field image segmentation method, which comprises the following steps:

[0006] determining a gradient magnitude of the preprocessed chest image and integrating the gradient magnitude;

[0007] Determine a thoracic image according to the integration result, and segment the thoracic image into a left thoracic image and a right thoracic image according to a left lung segmentation point and a right lung segmentation point;

[0008] performing costal margin segmentation, lung apex segmentation, and transverse and mediastinum segmentation on the left thoracic image and the right thoracic image, respectively, to obtain left and right costal margin segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinum segmentation images;

[0009] The apex region boundary, the transverse mediastinum region boundary and the costal margin region boundary are determined based on the left and right costal margin segmentation images, the left and right apex region segmentation images and the left and right transverse and mediastinum segmentation images, and the apex region boundary, the transverse mediastinum region boundary and the costal margin region boundary are connected to obtain a lung field image.

[0010] Optionally, the step of performing costal margin segmentation, lung apex segmentation, and transverse and mediastinal segmentation on the left thoracic image and the right thoracic image respectively to obtain left and right costal margin segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinal segmentation images includes:

[0011] performing noise reduction processing on the left thoracic image and the right thoracic image, and traversing the processed left thoracic image and the processed right thoracic image according to a preset directional derivative;

[0012] The traversed left chest image and the traversed right chest image are superimposed, and the superimposed left chest image and the superimposed right chest image are binarized to obtain left and right costal margin binary images;

[0013] Performing boundary detection on the left and right rib margin binary images to obtain left and right rib margin segmentation images;

[0014] performing image enhancement and binarization processing on the processed left thoracic image and the processed right thoracic image to obtain left and right transverse and mediastinal binary images;

[0015] Performing opening and closing operations and edge detection on the left and right transverse and mediastinal binary images to obtain left and right transverse and mediastinal edge binary images;

[0016] Determine the left and right transverse mediastinum segmentation images according to the left and right transverse mediastinum binary images and the left and right transverse mediastinum edge binary images;

[0017] The left and right pulmonary apex detection areas are determined according to the left and right costal margin segmentation images, and boundary detection is performed on the left and right pulmonary apex detection areas to obtain the left and right pulmonary apex segmentation images.

[0018] Optionally, the step of performing boundary detection on the left and right rib margin binary images to obtain left and right rib margin segmentation images includes:

[0019] Performing opening and closing operations on the left and right rib margin binary images, and determining the rib margin transverse gradient angle and the rib margin longitudinal gradient angle of the processed left and right rib margin binary images;

[0020] Performing an AND operation on the processed left and right costal margin binary images, the costal margin transverse gradient angle, and the costal margin longitudinal gradient angle;

[0021] Perform gradient angle screening on the left and right rib margin calculation results, and select the left and right rib margin connected domains based on the left and right rib margin angle screening results;

[0022] The left and right rib margin segmentation images are determined according to the target left and right rib margin connected regions.

[0023] Optionally, the step of determining the left and right transverse mediastinum segmentation images according to the left and right transverse mediastinum binary images and the left and right transverse mediastinum edge binary images includes:

[0024] Determine the transverse mediastinum transverse gradient angle and the transverse mediastinum longitudinal gradient angle according to the left and right transverse mediastinum binary images;

[0025] Performing an AND operation on the left and right transverse mediastinum edge binary images, the transverse mediastinum transverse gradient angle, and the transverse mediastinum longitudinal gradient angle;

[0026] Perform gradient angle screening on the left and right transverse and mediastinum calculation results, and select the left and right transverse and mediastinum connected domains according to the left and right transverse and mediastinum angle screening results;

[0027] The left and right transverse mediastinum segmentation images are determined according to the target left and right transverse mediastinum connected areas.

[0028] Optionally, the step of determining left and right lung apex detection areas based on the left and right costal margin segmentation images, and performing boundary detection on the left and right lung apex detection areas to obtain left and right lung apex segmentation images includes:

[0029] determining the coordinates of the left and right apexes of the costal margins according to the left and right costal margin segmentation images, and determining the coordinate points of the left and right lung apexes according to the processed left thorax image and the processed right thorax image;

[0030] Determine the left and right lung apex detection areas according to the coordinates of the left and right apexes of the costal margin and the coordinate points of the left and right lung apexes of the target;

[0031] Performing noise reduction and image enhancement processing on the left and right lung apex detection regions, and traversing the processed left and right lung apex detection regions according to the preset directional derivatives;

[0032] The left and right lung apex detection areas are superimposed after traversal, and the superimposed left and right lung apex detection areas are binarized to obtain left and right lung apex binary images;

[0033] The left and right lung apex binary images are fitted to obtain left and right lung apex segmentation images.

[0034] Optionally, the step of determining the gradient amplitude of the preprocessed chest image and integrating the gradient amplitude includes:

[0035] Determining a transverse gradient amplitude, a longitudinal gradient amplitude, and a total gradient amplitude of the preprocessed chest image;

[0036] The transverse gradient amplitude and the longitudinal gradient amplitude are vertically integrated, and the total gradient amplitude is transversely integrated and longitudinally integrated.

[0037] Optionally, the step of determining a thoracic image according to the integration result, and segmenting the thoracic image into a left thoracic image and a right thoracic image according to a left lung segmentation point and a right lung segmentation point includes:

[0038] determining a first ratio based on the transverse vertical integral and the transverse integral, and determining a second ratio based on the longitudinal vertical integral and the longitudinal integral;

[0039] When both the first ratio and the second ratio are less than a preset threshold, denoising the preprocessed chest image;

[0040] determining a local ratio maximum value according to the first ratio and the second ratio, and determining a thorax image according to the local ratio maximum value and the denoised thorax image;

[0041] Integrating the grayscale values ​​in the vertical direction of the thoracic image, and determining the left lung segmentation point and the right lung segmentation point according to the vertical grayscale value integration;

[0042] The thoracic image is segmented into a left thoracic image and a right thoracic image according to the left lung segmentation point and the right lung segmentation point.

[0043] Optionally, the step of determining the apex region boundary, the transverse mediastinum region boundary, and the costal margin region boundary based on the left and right costal margin segmentation images, the left and right apex region segmentation images, and the left and right transverse and mediastinum region segmentation images includes:

[0044] Determine a first apex Euclidean distance between the apex and the mediastinum according to the left and right apex segmentation images and the left and right transverse and mediastinal segmentation images, and determine a second apex Euclidean distance between the apex and the costal margin according to the left and right apex segmentation images and the left and right costal margin segmentation images;

[0045] determining a pulmonary apex region boundary according to a pulmonary apex point corresponding to the first pulmonary apex Euclidean distance and a pulmonary apex point corresponding to the second pulmonary apex Euclidean distance;

[0046] determining a first transverse-mediastinum Euclidean distance between the costal margin and the diaphragm according to the left and right costal margin segmentation images and the left and right transverse-mediastinum segmentation images;

[0047] determining a transverse mediastinum region boundary according to a transverse mediastinum endpoint corresponding to the first lung apex Euclidean distance and a transverse mediastinum endpoint corresponding to the first transverse mediastinum Euclidean distance;

[0048] The boundary of the costal margin region is determined according to the costal margin endpoint corresponding to the second lung apex Euclidean distance and the costal margin endpoint corresponding to the first transverse and mediastinal Euclidean distance.

[0049] In addition, to achieve the above-mentioned purpose, the present invention also proposes a lung field image segmentation device, which includes: an integral determination module and an image determination module;

[0050] The integral determination module is used to determine the gradient amplitude of the preprocessed chest image and integrate the gradient amplitude;

[0051] The image determination module is used to determine a thoracic image according to the integration result, and to segment the thoracic image into a left thoracic image and a right thoracic image according to a left lung segmentation point and a right lung segmentation point;

[0052] The image determination module is further configured to perform costal margin segmentation, lung apex segmentation, and transverse and mediastinal segmentation on the left thoracic image and the right thoracic image, respectively, to obtain left and right costal margin segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinal segmentation images;

[0053] The image determination module is further used to determine the apex region boundary, the transverse mediastinum region boundary and the costal margin region boundary based on the left and right costal margin segmentation images, the left and right apex region segmentation images and the left and right transverse and mediastinum segmentation images, and connect the apex region boundary, the transverse and mediastinum region boundary and the costal margin region boundary to obtain a lung field image.

[0054] In addition, to achieve the above-mentioned purpose, the present invention also proposes a storage medium, on which a lung field image segmentation program is stored. When the lung field image segmentation program is executed by a processor, the lung field image segmentation method described above is implemented.

[0055] The present invention discloses a lung field image segmentation method, device and storage medium. The method comprises: determining the gradient amplitude of a preprocessed chest image and integrating the gradient amplitude; determining a chest image according to the integration result, and dividing the chest image into a left chest image and a right chest image according to a left lung segmentation point and a right lung segmentation point; performing costal margin segmentation, lung apex segmentation and transverse and mediastinum segmentation on the left chest image and the right chest image respectively to obtain left and right costal margin segmentation images, left and right lung apex segmentation images and left and right transverse and mediastinum segmentation images; determining a lung apex region boundary, a transverse mediastinum region boundary and a costal margin region boundary according to the left and right costal margin segmentation images, the left and right lung apex segmentation images and the left and right transverse and mediastinum segmentation images, and connecting the lung apex region boundary, the transverse mediastinum region boundary and the costal margin region boundary to obtain a lung field image. The present invention integrates the chest image to remove unnecessary information to obtain left and right chest images, and performs costal margin segmentation, lung apex segmentation and transverse and mediastinal segmentation on the left and right chest images. Instead of segmenting the entire chest image, the present invention performs fine regional segmentation. Based on the segmented images, the left and right lung apex region boundaries, transverse and mediastinal region boundaries and costal margin region boundaries are determined and connected to obtain a lung field image, thereby improving the accuracy of lung field image segmentation. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 Schematic diagram of the process of the first embodiment of the lung field image segmentation method of the present invention;

[0057] Figure 2 Schematic diagram of the flow of the second embodiment of the lung field image segmentation method of the present invention;

[0058] Figure 3 The initial chest image of an embodiment of the lung field image segmentation method of the present invention;

[0059] Figure 4 A lung field segmentation map according to an embodiment of a lung field image segmentation method of the present invention;

[0060] Figure 5 The left costal margin segmentation image of the first embodiment of the lung field image segmentation method of the present invention;

[0061] Figure 6 The left transverse and mediastinal segmentation image of the first embodiment of the lung field image segmentation method of the present invention;

[0062] Figure 7 The left lung apex segmentation image of the first embodiment of the lung field image segmentation method of the present invention;

[0063] Figure 8 This is a structural block diagram of the first embodiment of the lung field image segmentation device of the present invention.

[0064] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0065] Reference Figure 1 , Figure 1 FIG1 is a flow chart of the first embodiment of the lung field image segmentation method of the present invention, which provides the first embodiment of the lung field image segmentation method of the present invention.

[0066] Step S10: determining the gradient magnitude of the preprocessed chest image and integrating the gradient magnitude.

[0067] It should be noted that the execution subject of this embodiment can be a computer software service device with data processing, network communication and program running functions, such as a lung field image segmentation device, or other electronic devices that can achieve the same or similar functions. This embodiment does not impose any restrictions on this.

[0068] It is understandable that the preprocessing of the chest image may include performing low-pass filtering on the chest image, reducing the filtered image to speed up image processing, and performing logarithmic transformation on the image to enhance the image.

[0069] It should be noted that the transverse and longitudinal gradient amplitudes of the pre-processed chest image may be determined, and the transverse gradient amplitude and the longitudinal gradient amplitude may be integrated.

[0070] Furthermore, in order to improve the accuracy of lung field image segmentation, step S10 in this embodiment may include:

[0071] Determining a transverse gradient amplitude, a longitudinal gradient amplitude, and a total gradient amplitude of the preprocessed chest image;

[0072] The transverse gradient amplitude and the longitudinal gradient amplitude are vertically integrated, and the total gradient amplitude is transversely integrated and longitudinally integrated.

[0073] In a specific implementation, the transverse gradient amplitude, the longitudinal gradient amplitude and the total gradient amplitude of the preprocessed chest image are calculated, and the vertical integral of the transverse gradient amplitude, the vertical integral of the longitudinal gradient amplitude and the transverse integral and longitudinal integral of the total gradient amplitude are calculated.

[0074] Step S20: determining a thoracic image according to the integration result, and segmenting the thoracic image into a left thoracic image and a right thoracic image according to the left lung segmentation point and the right lung segmentation point.

[0075] It should be noted that the ratio is calculated based on the integration result and the local ratio maximum is determined based on the ratio calculation result. The chest image is obtained based on the local ratio maximum and the characteristics of the chest. Information that does not belong to the chest image needs to be removed, such as some blank background information after the arms and head.

[0076] It should be noted that the grayscale values ​​of the chest image in the vertical direction are integrated, and the maximum value of the grayscale value integral sum is determined. The division points of the left and right lungs of the left chest image and the right chest image are determined based on the maximum value, and the chest image is divided into the left chest image and the right chest image based on the division points of the left and right lungs.

[0077] Furthermore, in order to improve the convenience of lung field image segmentation, step S20 in this embodiment may include:

[0078] determining a first ratio based on the transverse vertical integral and the transverse integral, and determining a second ratio based on the longitudinal vertical integral and the longitudinal integral;

[0079] When both the first ratio and the second ratio are less than a preset threshold, denoising the preprocessed chest image;

[0080] determining a local ratio maximum value according to the first ratio and the second ratio, and determining a thorax image according to the local ratio maximum value and the denoised thorax image;

[0081] Integrating the grayscale values ​​in the vertical direction of the thoracic image, and determining the left lung segmentation point and the right lung segmentation point according to the vertical grayscale value integration;

[0082] The thoracic image is segmented into a left thoracic image and a right thoracic image according to the left lung segmentation point and the right lung segmentation point.

[0083] It should be noted that the calculated ratio is compared with the preset threshold for noise processing, that is, only when both the first ratio and the second ratio are smaller than the preset threshold, the pixels in the direction corresponding to the integral of the first ratio and the second ratio are discarded for noise.

[0084] It can be understood that the maximum integral value is determined by the sum of the grayscale value integrals in the vertical direction of the branch, and the left lung segmentation point and the right lung segmentation point are determined based on the maximum integral value.

[0085] Step S30: performing costal margin segmentation, lung apex segmentation, and transverse and mediastinal segmentation on the left thoracic image and the right thoracic image respectively to obtain left and right costal margin segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinal segmentation images.

[0086] It should be noted that the rib margin segmentation for the left chest image and the right chest image can be performed by first traversing the left and right chest images according to the template of the directional derivative, superimposing the traversal results and binarizing the superimposed results to obtain a rib margin binary image, performing opening and closing operations on the rib margin binary image and calculating the gradient angle of each pixel in the superimposed rib margin chest image, performing an AND operation on the rib margin binary image after the operation and the gradient angle of each pixel and filtering the operation results, and determining the maximum connected domain based on the screening results. The maximum connected domain is the left and right rib margin segmentation image.

[0087] It should be noted that the apex segmentation can be performed by constructing a apex detection area based on the coordinates of each point in the left and right costal margin segmentation images, performing noise reduction and image enhancement processing on the apex detection area, traversing the processed apex detection area image according to the directional derivative template, superimposing the traversal results and binarizing the superimposed results to obtain a apex binary image, performing parameter fitting on the apex binary image, and obtaining left and right apex segmentation images.

[0088] It should be noted that the transverse and mediastinum segmentation can be performed by performing image enhancement and binarization processing on the left and right chest images to obtain a transverse and mediastinum binary image, performing opening and closing operations and edge detection on the transverse and mediastinum binary image to obtain a transverse and mediastinum edge binary image, calculating the gradient angle of each pixel in the transverse and mediastinum edge binary image, performing an AND operation on the transverse and mediastinum edge binary image and the gradient angle of each pixel, and performing angle screening on the operation results, and determining the maximum connected domain based on the screening results. The maximum connected domain is the left and right transverse and mediastinum segmentation images.

[0089] Step S40: Determine the apex region boundary, the transverse mediastinum region boundary, and the costal margin region boundary based on the left and right costal margin segmentation images, the left and right apex region segmentation images, and the left and right transverse and mediastinum region segmentation images, and connect the apex region boundary, the transverse and mediastinum region boundary, and the costal margin region boundary to obtain a lung field image.

[0090] It should be noted that the two shortest Euclidean points between the costal margin, the apex of the lung and the transverse mediastinum are calculated as the costal margin segmentation point, the apex of the lung segmentation point and the transverse mediastinum segmentation point, respectively. The apex of the lung region boundary, the transverse mediastinum of the lung region boundary and the costal margin of the lung region boundary are determined according to the costal margin segmentation point, the apex of the lung region boundary and the transverse mediastinum of the lung region boundary, and the costal margin of the lung region boundary are connected to obtain a lung field image.

[0091] Furthermore, in order to improve the accuracy of lung field image segmentation, step S40 in this embodiment may include:

[0092] Determine a first apex Euclidean distance between the apex and the mediastinum according to the left and right apex segmentation images and the left and right transverse and mediastinal segmentation images, and determine a second apex Euclidean distance between the apex and the costal margin according to the left and right apex segmentation images and the left and right costal margin segmentation images;

[0093] determining a pulmonary apex region boundary according to a pulmonary apex point corresponding to the first pulmonary apex Euclidean distance and a pulmonary apex point corresponding to the second pulmonary apex Euclidean distance;

[0094] determining a first transverse-mediastinum Euclidean distance between the costal margin and the diaphragm according to the left and right costal margin segmentation images and the left and right transverse-mediastinum segmentation images;

[0095] determining a transverse mediastinum region boundary according to a transverse mediastinum endpoint corresponding to the first lung apex Euclidean distance and a transverse mediastinum endpoint corresponding to the first transverse mediastinum Euclidean distance;

[0096] The boundary of the costal margin region is determined according to the costal margin endpoint corresponding to the second lung apex Euclidean distance and the costal margin endpoint corresponding to the first transverse and mediastinal Euclidean distance.

[0097] It should be noted that the first apical Euclidean distance is determined based on the two shortest Euclidean points between the apex and the mediastinum. Similarly, the second apical Euclidean distance is determined based on the two shortest Euclidean points between the apex and the costal margin, and the first transverse mediastinal Euclidean distance is determined based on the two shortest Euclidean points between the costal margin and the diaphragm.

[0098] For ease of understanding, refer to Figure 3 and Figure 4 To explain, Figure 3 is the initial chest image, Figure 4 is the lung field segmentation map, from Figure 3 and Figure 4 It can be clearly seen that the lung field segmentation performed by this embodiment can very accurately segment the entire lung field.

[0099] This embodiment determines the gradient amplitude of the preprocessed chest image and integrates the gradient amplitude; determines a chest image based on the integration result, and segments the chest image into a left chest image and a right chest image based on the left lung segmentation point and the right lung segmentation point; performs costal margin segmentation, pulmonary apex segmentation, and transverse and mediastinal segmentation on the left chest image and the right chest image, respectively, to obtain left and right costal margin segmentation images, left and right pulmonary apex segmentation images, and left and right transverse and mediastinal segmentation images; determines the pulmonary apex region boundary, the transverse mediastinum region boundary, and the costal margin region boundary based on the left and right costal margin segmentation images, the left and right pulmonary apex segmentation images, and the left and right transverse and mediastinum segmentation images, and connects the pulmonary apex region boundary, the transverse mediastinum region boundary, and the costal margin region boundary to obtain a lung field image. In this embodiment, the chest image is integrated to remove unnecessary information to obtain left and right chest images, and the left and right chest images are segmented by costal margin, lung apex, and transverse and mediastinum. Instead of segmenting the entire chest image, a fine segmentation of regions is performed. Based on the segmented images, the left and right lung apex region boundaries, transverse and mediastinal region boundaries, and costal margin region boundaries are determined and connected to obtain a lung field image, thereby improving the accuracy of lung field image segmentation.

[0100] Reference Figure 2 , Figure 2 This is a flow chart of the second embodiment of the lung field image segmentation method of the present invention. Figure 1 The first embodiment shown here provides a second embodiment of the lung field image segmentation method of the present invention.

[0101] In the second embodiment, step S30 includes:

[0102] Step S301: performing noise reduction processing on the left thoracic image and the right thoracic image, and traversing the processed left thoracic image and the processed right thoracic image according to a preset directional derivative.

[0103] It can be understood that the large-scale Gaussian blurring is performed on the left thorax image and the right thorax image to reduce the detail information of the left thorax image and the right thorax image.

[0104] It should be noted that it is necessary to construct a preset directional derivative template and set the weighted depth of the directional derivative template. The directional derivative can be:

[0105]

[0106] Where f(x0,y0) represents the directional derivative, l represents the unit vector in the direction, cosa represents and cosβ represent the direction cosines.

[0107] It should be noted that all directions of each pixel in the left and right rib margin contours after denoising are traversed according to the template of the directional derivative.

[0108] Step S302: superimposing the traversed left thoracic image and the traversed right thoracic image, and performing binarization processing on the superimposed left thoracic image and the superimposed right thoracic image to obtain left and right costal margin binary images.

[0109] It should be noted that the traversed left chest image and the traversed right chest image are superimposed on the left chest image and the right chest image before traversal.

[0110] It can be understood that the left and right rib cage images are binarized using the maximum inter-class variance method to obtain binary images of the left and right costal margins.

[0111] Step S303: performing boundary detection on the left and right rib margin binary images to obtain left and right rib margin segmentation images.

[0112] It should be noted that the morphological opening and closing operations and refinement processing are performed on the left and right rib margin binary images, and the horizontal gradient calculation and the vertical gradient calculation are performed on the superimposed left thoracic image and the superimposed right thoracic image to obtain the gradient angle of each pixel. The gradient angle and the morphologically processed left and right rib margin binary images are ANDed, and the results of the AND operation are angle-screened. The largest connected domain is determined based on the screened gradient angle to obtain the left and right rib margin segmentation images.

[0113] Furthermore, in order to improve the accuracy of lung field image segmentation, step S203 in this embodiment may include:

[0114] Performing opening and closing operations on the left and right rib margin binary images, and determining the rib margin transverse gradient angle and the rib margin longitudinal gradient angle of the processed left and right rib margin binary images;

[0115] Performing an AND operation on the processed left and right costal margin binary images, the costal margin transverse gradient angle, and the costal margin longitudinal gradient angle;

[0116] Perform gradient angle screening on the left and right rib margin calculation results, and select the left and right rib margin connected domains based on the left and right rib margin angle screening results;

[0117] The left and right rib margin segmentation images are determined according to the target left and right rib margin connected regions.

[0118] It should be understood that the target left and right rib margin connected domain is the largest left and right rib margin connected domain.

[0119] For ease of understanding, refer to Figure 5 To explain, Figure 5 This is the segmentation image of the left costal margin. The bold part of the line in the figure is the segmented left costal margin.

[0120] Step S304: performing image enhancement and binarization processing on the processed left thoracic image and the processed right thoracic image to obtain left and right transverse and mediastinal binary images.

[0121] It can be understood that performing image enhancement and binarization processing may be performing contrast enhancement and maximum inter-class variance binarization processing on the left thoracic image after noise reduction processing and the right thoracic image after processing.

[0122] Step S305: performing opening and closing operations and edge detection on the left and right transverse and mediastinal binary images to obtain left and right transverse and mediastinal edge binary images.

[0123] Step S306: determining left and right transverse mediastinum segmentation images according to the left and right transverse mediastinum binary images and the left and right transverse mediastinum edge binary images.

[0124] It should be noted that the horizontal and vertical gradients are calculated based on the left and right transverse and mediastinal binary images to determine the gradient angle of each pixel. The gradient angle of each pixel is ANDed with the left and right transverse and mediastinal edge binary images. The gradient angles are screened based on the calculation results and the characteristics of the diaphragm and mediastinum. The largest connected domain is selected based on the screened gradient angles to obtain the left and right transverse and mediastinal segmentation images.

[0125] Furthermore, in order to improve the accuracy of segmentation of the lung field image, step S306 of this embodiment may include:

[0126] Determine the transverse mediastinum transverse gradient angle and the transverse mediastinum longitudinal gradient angle according to the left and right transverse mediastinum binary images;

[0127] Performing an AND operation on the left and right transverse mediastinum edge binary images, the transverse mediastinum transverse gradient angle, and the transverse mediastinum longitudinal gradient angle;

[0128] Perform gradient angle screening on the left and right transverse and mediastinum calculation results, and select the left and right transverse and mediastinum connected domains according to the left and right transverse and mediastinum angle screening results;

[0129] The left and right transverse mediastinum segmentation images are determined according to the target left and right transverse mediastinum connected areas.

[0130] It can be understood that the target left and right transverse and mediastinal connected domain is the largest left and right transverse and mediastinal connected domain.

[0131] For ease of understanding, refer to Figure 6 To explain, Figure 6 This is the segmentation image of the left transverse mediastinum. The bold part in the figure is the segmented left transverse mediastinum.

[0132] Step S307: determining left and right lung apex detection regions according to the left and right costal margin segmentation images, and performing boundary detection on the left and right lung apex detection regions to obtain left and right lung apex segmentation images.

[0133] It should be noted that the rectangular area formed by the oblique sides composed of the coordinate points of the rib edge vertices in the left and right rib edge segmentation images and the coordinate points of the upper left and upper right corners in the left and right thoracic images is the left and right lung apex detection area.

[0134] Furthermore, in order to improve the accuracy of lung field image segmentation, step S307 in this embodiment may include:

[0135] determining the coordinates of the left and right apexes of the costal margins according to the left and right costal margin segmentation images, and determining the coordinate points of the left and right lung apexes according to the processed left thorax image and the processed right thorax image;

[0136] Determine the left and right lung apex detection areas according to the coordinates of the left and right apexes of the costal margin and the coordinate points of the left and right lung apexes of the target;

[0137] Performing noise reduction and image enhancement processing on the left and right lung apex detection regions, and traversing the processed left and right lung apex detection regions according to the preset directional derivatives;

[0138] The left and right lung apex detection areas are superimposed after traversal, and the superimposed left and right lung apex detection areas are binarized to obtain left and right lung apex binary images;

[0139] The left and right lung apex binary images are fitted to obtain left and right lung apex segmentation images.

[0140] It should be noted that, according to the features of the lung apex edge and the left and right flying sword binary images, the Hough space parameter fitting of the quadratic function is performed to obtain the fitted left and right lung apex segmentation images.

[0141] For ease of understanding, refer to Figure 7 To explain, Figure 7 This is the segmentation image of the left pulmonary apex. The bold part in the figure is the segmented left pulmonary apex.

[0142] This embodiment superimposes the traversed left chest image and the traversed right chest image, and performs binarization processing on the superimposed left chest image and the superimposed right chest image to obtain left and right costal margin binary images; performs boundary detection on the left and right costal margin binary images to obtain left and right costal margin segmentation images; performs image enhancement and binarization processing on the processed left chest image and the processed right chest image to obtain left and right transverse and mediastinum binary images; performs opening and closing operations and edge detection on the left and right transverse and mediastinum binary images to obtain left and right transverse and mediastinum edge binary images; determines left and right transverse and mediastinum segmentation images based on the left and right transverse and mediastinum binary images and the left and right transverse and mediastinum edge binary images; determines left and right lung apex detection areas based on the left and right costal margin segmentation images, and performs boundary detection on the left and right lung apex detection areas to obtain left and right lung apex segmentation images. This embodiment segments the costal margins, lung apex, and transverse and mediastinum of the left and right thoracic images respectively to obtain left and right costal margin binary images, left and right transverse and mediastinum segmentation images, and left and right lung apex segmentation images, thereby eliminating the need for manual labeling and improving the rate of lung field image segmentation.

[0143] In addition, an embodiment of the present invention further provides a storage medium on which a lung field image segmentation program is stored. When the lung field image segmentation program is executed by a processor, the lung field image segmentation method described above is implemented.

[0144] In addition, refer to Figure 8 , an embodiment of the present invention further provides a lung field image segmentation device, the lung field image segmentation device comprising: an integral determination module 10 and an image determination module 20;

[0145] The integral determination module 10 is used to determine the gradient amplitude of the preprocessed chest image and integrate the gradient amplitude;

[0146] The image determination module 20 is used to determine a thoracic image according to the integration result, and to segment the thoracic image into a left thoracic image and a right thoracic image according to a left lung segmentation point and a right lung segmentation point;

[0147] The image determination module 20 is further configured to perform costal segmentation, lung apex segmentation, and transverse and mediastinal segmentation on the left thoracic image and the right thoracic image, respectively, to obtain left and right costal segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinal segmentation images;

[0148] The image determination module 20 is further used to determine the apex region boundary, the transverse mediastinum region boundary and the costal margin region boundary based on the left and right costal margin segmentation images, the left and right apex region segmentation images and the left and right transverse and mediastinum segmentation images, and connect the apex region boundary, the transverse and mediastinum region boundary and the costal margin region boundary to obtain a lung field image.

[0149] This embodiment determines the gradient amplitude of the preprocessed chest image and integrates the gradient amplitude; determines a chest image based on the integration result, and segments the chest image into a left chest image and a right chest image based on the left lung segmentation point and the right lung segmentation point; performs costal margin segmentation, pulmonary apex segmentation, and transverse and mediastinal segmentation on the left chest image and the right chest image, respectively, to obtain left and right costal margin segmentation images, left and right pulmonary apex segmentation images, and left and right transverse and mediastinal segmentation images; determines the pulmonary apex region boundary, the transverse mediastinum region boundary, and the costal margin region boundary based on the left and right costal margin segmentation images, the left and right pulmonary apex segmentation images, and the left and right transverse and mediastinum segmentation images, and connects the pulmonary apex region boundary, the transverse mediastinum region boundary, and the costal margin region boundary to obtain a lung field image. In this embodiment, the chest image is integrated to remove unnecessary information to obtain left and right chest images, and the left and right chest images are segmented by costal margin, lung apex, and transverse and mediastinum. Instead of segmenting the entire chest image, a fine segmentation of regions is performed. Based on the segmented images, the left and right lung apex region boundaries, transverse and mediastinal region boundaries, and costal margin region boundaries are determined and connected to obtain a lung field image, thereby improving the accuracy of lung field image segmentation.

[0150] Based on the first embodiment of the lung field image segmentation device of the present invention, a second embodiment of the lung field image segmentation device of the present invention is proposed.

[0151] In this embodiment, the image determination module 20 is configured to perform noise reduction processing on the left chest image and the right chest image, and traverse the processed left chest image and the processed right chest image according to preset directional derivatives.

[0152] Furthermore, the image determination module 20 is further configured to superimpose the traversed left chest image and the traversed right chest image, and perform binarization processing on the superimposed left chest image and the superimposed right chest image to obtain left and right costal margin binary images.

[0153] Furthermore, the image determination module 20 is further configured to perform boundary detection on the left and right rib margin binary images to obtain left and right rib margin segmentation images.

[0154] Furthermore, the image determination module 20 is further configured to perform image enhancement and binarization processing on the processed left thoracic image and the processed right thoracic image to obtain left and right transverse and mediastinal binary images.

[0155] Furthermore, the image determination module 20 is further configured to perform opening and closing operations and edge detection on the left and right transverse and mediastinal binary images to obtain left and right transverse and mediastinal edge binary images.

[0156] Furthermore, the image determination module 20 is further configured to determine left and right transverse mediastinum segmentation images according to the left and right transverse mediastinum binary images and the left and right transverse mediastinum edge binary images.

[0157] Furthermore, the image determination module 20 is further configured to determine left and right pulmonary apex detection regions based on the left and right costal margin segmentation images, and perform boundary detection on the left and right pulmonary apex detection regions to obtain left and right pulmonary apex segmentation images.

[0158] Furthermore, the image determination module 20 is further configured to perform opening and closing operations on the left and right rib margin binary images, and determine the rib margin transverse gradient angle and the rib margin longitudinal gradient angle of the processed left and right rib margin binary images.

[0159] Furthermore, the image determination module 20 is further configured to perform an AND operation on the processed left and right rib margin binary images, the rib margin transverse gradient angle, and the rib margin longitudinal gradient angle.

[0160] Furthermore, the image determination module 20 is further configured to perform gradient angle screening on the left and right rib margin calculation results, and select the left and right rib margin connected domains according to the left and right rib margin angle screening results.

[0161] Furthermore, the image determination module 20 is further configured to determine the left and right rib margin segmentation images according to the target left and right rib margin connected domains.

[0162] Furthermore, the image determination module 20 is further configured to determine a transverse gradient angle of the transverse mediastinum and a longitudinal gradient angle of the transverse mediastinum according to the left and right transverse and mediastinal binary images.

[0163] Furthermore, the image determination module 20 is further configured to perform an AND operation on the left and right transverse mediastinum edge binary images, the transverse mediastinum transverse gradient angle, and the transverse mediastinum longitudinal gradient angle.

[0164] Furthermore, the image determination module 20 is further configured to perform gradient angle screening on the left and right transverse and mediastinum calculation results, and select the left and right transverse and mediastinum connected domains according to the left and right transverse and mediastinum angle screening results.

[0165] Furthermore, the image determination module 20 is further configured to determine left and right transverse and mediastinum segmentation images according to the target left and right transverse and mediastinum connected domains.

[0166] Furthermore, the image determination module 20 is further configured to determine the coordinates of the left and right apexes of the costal margins based on the left and right costal margin segmentation images, and to determine the coordinate points of the target left and right lung apexes based on the processed left thoracic image and the processed right thoracic image.

[0167] Furthermore, the image determination module 20 is further configured to determine left and right lung apex detection areas according to the left and right apex coordinates of the costal margin and the left and right lung apex coordinate points of the target.

[0168] Furthermore, the image determination module 20 is further configured to perform noise reduction and image enhancement processing on the left and right lung apex detection regions, and traverse the processed left and right lung apex detection regions according to the preset directional derivatives.

[0169] Furthermore, the image determination module 20 is further configured to superimpose the traversed left and right lung apex detection regions, and perform binarization processing on the superimposed left and right lung apex detection regions to obtain left and right lung apex binary images.

[0170] Furthermore, the image determination module 20 is further configured to fit the left and right lung apex binary images to obtain left and right lung apex segmentation images.

[0171] Furthermore, the integral determination module 10 is further configured to determine a transverse gradient amplitude, a longitudinal gradient amplitude, and a total gradient amplitude of the preprocessed chest image.

[0172] Furthermore, the integral determination module 10 is further configured to perform vertical integration on the transverse gradient amplitude and the longitudinal gradient amplitude, and perform transverse integration and longitudinal integration on the total gradient amplitude.

[0173] Furthermore, the image determination module 20 is further configured to determine a first ratio according to the horizontal vertical integral and the horizontal integral, and to determine a second ratio according to the vertical vertical integral and the vertical integral.

[0174] Furthermore, the image determination module 20 is further configured to perform denoising on the preprocessed chest image when both the first ratio and the second ratio are smaller than a preset threshold.

[0175] Furthermore, the image determination module 20 is further configured to determine a local ratio maximum value according to the first ratio and the second ratio, and determine a thorax image according to the local ratio maximum value and the denoised chest image.

[0176] Furthermore, the image determination module 20 is further configured to integrate the grayscale value in the vertical direction of the thoracic image, and determine the left lung segmentation point and the right lung segmentation point according to the vertical grayscale value integral and the local ratio maximum.

[0177] Furthermore, the image determination module 20 is further configured to segment the thoracic image into a left thoracic image and a right thoracic image according to the left lung segmentation point and the right lung segmentation point.

[0178] Furthermore, the image determination module 20 is also used to determine a first apex Euclidean distance between the apex and the mediastinum based on the left and right apex segmentation images and the left and right transverse and mediastinal segmentation images, and to determine a second apex Euclidean distance between the apex and the costal margin based on the left and right apex segmentation images and the left and right costal margin segmentation images.

[0179] Furthermore, the image determination module 20 is further configured to determine a boundary of a lung apex region according to a lung apex point corresponding to the first lung apex Euclidean distance and a lung apex point corresponding to the second lung apex Euclidean distance.

[0180] Furthermore, the image determination module 20 is further configured to determine a second transverse-mediastinum Euclidean distance between the costal margin and the diaphragm according to the left and right costal margin segmentation images and the left and right transverse-mediastinum segmentation images.

[0181] Furthermore, the image determination module 20 is further configured to determine a transverse mediastinum region boundary according to a transverse mediastinum endpoint corresponding to the first lung apex Euclidean distance and a transverse mediastinum endpoint corresponding to the second transverse mediastinum Euclidean distance.

[0182] Furthermore, the image determination module 20 is further configured to determine a costal margin region boundary according to a costal margin endpoint corresponding to the second apex Euclidean distance and a costal margin endpoint corresponding to the second transverse and mediastinal Euclidean distance.

[0183] Other embodiments or specific implementations of the lung field image segmentation device of the present invention can refer to the above-mentioned method embodiments and will not be repeated here.

[0184] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0185] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0186] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, or of course by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory image (ROM) / random access memory (RAM), a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0187] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A lung field image segmentation method, characterized in that: The lung field image segmentation method comprises the following steps: determining a gradient magnitude of the preprocessed chest image and integrating the gradient magnitude; Determine a thoracic image according to the integration result, and segment the thoracic image into a left thoracic image and a right thoracic image according to a left lung segmentation point and a right lung segmentation point; performing costal margin segmentation, lung apex segmentation, and transverse and mediastinum segmentation on the left thoracic image and the right thoracic image, respectively, to obtain left and right costal margin segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinum segmentation images; determining a lung apex region boundary, a transverse mediastinum region boundary, and a costal margin region boundary according to the left and right costal margin segmentation images, the left and right lung apex segmentation images, and the left and right transverse mediastinum segmentation images, and connecting the lung apex region boundary, the transverse mediastinum region boundary, and the costal margin region boundary to obtain a lung field image; The step of performing costal margin segmentation, lung apex segmentation, and transverse and mediastinal segmentation on the left thoracic image and the right thoracic image respectively to obtain left and right costal margin segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinal segmentation images comprises: performing noise reduction processing on the left thoracic image and the right thoracic image, and traversing the processed left thoracic image and the processed right thoracic image according to a preset directional derivative; The traversed left chest image and the traversed right chest image are superimposed, and the superimposed left chest image and the superimposed right chest image are binarized to obtain left and right costal margin binary images; Performing boundary detection on the left and right rib margin binary images to obtain left and right rib margin segmentation images; performing image enhancement and binarization processing on the processed left thoracic image and the processed right thoracic image to obtain left and right transverse and mediastinal binary images; Performing opening and closing operations and edge detection on the left and right transverse and mediastinal binary images to obtain left and right transverse and mediastinal edge binary images; Determine the left and right transverse mediastinum segmentation images according to the left and right transverse mediastinum binary images and the left and right transverse mediastinum edge binary images; The left and right pulmonary apex detection areas are determined according to the left and right costal margin segmentation images, and boundary detection is performed on the left and right pulmonary apex detection areas to obtain the left and right pulmonary apex segmentation images.

2. The lung field image segmentation method according to claim 1, wherein: The step of performing boundary detection on the left and right rib margin binary images to obtain left and right rib margin segmentation images includes: Performing opening and closing operations on the left and right rib margin binary images, and determining the rib margin transverse gradient angle and the rib margin longitudinal gradient angle of the processed left and right rib margin binary images; Performing an AND operation on the processed left and right costal margin binary images, the costal margin transverse gradient angle, and the costal margin longitudinal gradient angle; Perform gradient angle screening on the left and right rib margin calculation results, and select the left and right rib margin connected domains based on the left and right rib margin angle screening results; The left and right rib margin segmentation images are determined according to the target left and right rib margin connected regions.

3. The lung field image segmentation method according to claim 1, wherein: The step of determining the left and right transverse mediastinum segmentation images according to the left and right transverse mediastinum binary images and the left and right transverse mediastinum edge binary images comprises: Determine the transverse mediastinum transverse gradient angle and the transverse mediastinum longitudinal gradient angle according to the left and right transverse mediastinum binary images; Performing an AND operation on the left and right transverse mediastinum edge binary images, the transverse mediastinum transverse gradient angle, and the transverse mediastinum longitudinal gradient angle; Perform gradient angle screening on the left and right transverse and mediastinum calculation results, and select the left and right transverse and mediastinum connected domains according to the left and right transverse and mediastinum angle screening results; The left and right transverse mediastinum segmentation images are determined according to the target left and right transverse mediastinum connected areas.

4. The lung field image segmentation method according to claim 1, wherein: The step of determining left and right lung apex detection areas based on the left and right costal margin segmentation images, and performing boundary detection on the left and right lung apex detection areas to obtain left and right lung apex segmentation images includes: determining the coordinates of the left and right apexes of the costal margins according to the left and right costal margin segmentation images, and determining the coordinate points of the left and right lung apexes according to the processed left thorax image and the processed right thorax image; Determine the left and right lung apex detection areas according to the coordinates of the left and right apexes of the costal margin and the coordinate points of the left and right lung apexes of the target; Performing noise reduction and image enhancement processing on the left and right lung apex detection regions, and traversing the processed left and right lung apex detection regions according to the preset directional derivatives; The left and right lung apex detection areas are superimposed after traversal, and the superimposed left and right lung apex detection areas are binarized to obtain left and right lung apex binary images; The left and right lung apex binary images are fitted to obtain left and right lung apex segmentation images.

5. The lung field image segmentation method according to claim 1, wherein: The step of determining the gradient amplitude of the preprocessed chest image and integrating the gradient amplitude comprises: Determining a transverse gradient amplitude, a longitudinal gradient amplitude, and a total gradient amplitude of the preprocessed chest image; The transverse gradient amplitude and the longitudinal gradient amplitude are vertically integrated, and the total gradient amplitude is transversely integrated and longitudinally integrated.

6. The lung field image segmentation method according to claim 1, wherein: The step of determining a thoracic image according to the integration result, and segmenting the thoracic image into a left thoracic image and a right thoracic image according to a left lung segmentation point and a right lung segmentation point comprises: determining a first ratio based on the transverse vertical integral and the transverse integral, and determining a second ratio based on the longitudinal vertical integral and the longitudinal integral; When both the first ratio and the second ratio are less than a preset threshold, denoising the preprocessed chest image; determining a local ratio maximum value according to the first ratio and the second ratio, and determining a thorax image according to the local ratio maximum value and the denoised thorax image; Integrating the grayscale values ​​in the vertical direction of the thoracic image, and determining the left lung segmentation point and the right lung segmentation point according to the vertical grayscale value integration; The thoracic image is segmented into a left thoracic image and a right thoracic image according to the left lung segmentation point and the right lung segmentation point.

7. The lung field image segmentation method according to claim 1, wherein: The step of determining the apex region boundary, the transverse mediastinum region boundary, and the costal margin region boundary based on the left and right costal margin segmentation images, the left and right apex region segmentation images, and the left and right transverse and mediastinum region segmentation images comprises: Determine a first apex Euclidean distance between the apex and the mediastinum according to the left and right apex segmentation images and the left and right transverse and mediastinal segmentation images, and determine a second apex Euclidean distance between the apex and the costal margin according to the left and right apex segmentation images and the left and right costal margin segmentation images; determining a pulmonary apex region boundary according to a pulmonary apex point corresponding to the first pulmonary apex Euclidean distance and a pulmonary apex point corresponding to the second pulmonary apex Euclidean distance; a step of determining a first transverse-mediastinum Euclidean distance between the costal margin and the diaphragm according to the left and right costal margin segmentation images and the left and right transverse-mediastinum segmentation images; determining a transverse mediastinum region boundary according to a transverse mediastinum endpoint corresponding to the first lung apex Euclidean distance and a transverse mediastinum endpoint corresponding to the first transverse mediastinum Euclidean distance; The boundary of the costal margin region is determined according to the costal margin endpoint corresponding to the second lung apex Euclidean distance and the costal margin endpoint corresponding to the first transverse and mediastinal Euclidean distance.

8. A lung field image segmentation device, characterized in that: The lung field image segmentation device includes: an integral determination module and an image determination module; The integral determination module is used to determine the gradient amplitude of the preprocessed chest image and integrate the gradient amplitude; The image determination module is used to determine a thoracic image according to the integration result, and to segment the thoracic image into a left thoracic image and a right thoracic image according to a left lung segmentation point and a right lung segmentation point; The image determination module is further configured to perform costal margin segmentation, lung apex segmentation, and transverse and mediastinal segmentation on the left thoracic image and the right thoracic image, respectively, to obtain left and right costal margin segmentation images, left and right lung apex segmentation images, and left and right transverse and mediastinal segmentation images; The image determination module is further configured to determine a lung apex region boundary, a transverse mediastinum region boundary, and a costal margin region boundary based on the left and right costal margin segmentation images, the left and right lung apex segmentation images, and the left and right transverse and mediastinum segmentation images, and connect the lung apex region boundary, the transverse and mediastinum region boundary, and the costal margin region boundary to obtain a lung field image; The image determination module is further configured to perform noise reduction processing on the left chest image and the right chest image, and traverse the processed left chest image and the processed right chest image according to a preset directional derivative; superimpose the traversed left chest image and the traversed right chest image, and perform binarization processing on the superimposed left chest image and the superimposed right chest image to obtain left and right rib margin binary images; perform boundary detection on the left and right rib margin binary images to obtain left and right rib margin segmentation images; perform image enhancement and binarization processing on the processed left chest image and the processed right chest image to obtain left and right transverse and mediastinum binary images; perform opening and closing operations and edge detection on the left and right transverse and mediastinum binary images to obtain left and right transverse and mediastinum edge binary images; determine left and right transverse and mediastinum segmentation images based on the left and right transverse and mediastinum binary images and the left and right transverse and mediastinum edge binary images; determine left and right lung apex detection areas based on the left and right rib margin segmentation images, and perform boundary detection on the left and right lung apex detection areas to obtain left and right lung apex segmentation images.

9. A storage medium, characterized in that: The storage medium stores a lung field image segmentation program, which, when executed by a processor, implements the steps of the lung field image segmentation method according to any one of claims 1 to 7.

Images