Ultrasonic image gallbladder form automatic identification method, system and equipment

Abstract

The invention relates to the technical field of ultrasonic image recognition, and discloses an ultrasonic image gall bladder form automatic recognition method, system and equipment, and the method comprises the steps: segmenting a gall bladder region from an ultrasonic image through a deep learning segmentation model, and extracting a central skeleton line through an image skeleton extraction algorithm; according to the position relation between two preset points on the central skeleton line and all contour points on the gallbladder area, determining two end points of a long axis and further determining a target long axis; and taking a line segment between any two contour points on the gall bladder area as a target line segment, and taking a target line segment which is vertical to the target long axis and corresponds to the maximum width value in each target line segment as a target short axis. According to the method, the image segmentation technology and the target extraction algorithm are added, the form of the gall bladder can be automatically recognized, the long axis and the short axis of the gall bladder are automatically recognized according to the form of the gall bladder in the ultrasonic image, and the recognition accuracy of the longest axis and the shortest axis of the gall bladder is improved when the form of the gall bladder is complex or irregular.

Description

Background Art

[0002] The gallbladder is an important part of the human digestive system, so the morphological analysis of the gallbladder is of great research significance. Ultrasound examination has the advantages of being economical, non-destructive, real-time, and portable, and is an important means of detecting gallbladder morphology. During gallbladder ultrasound examination, ultrasound doctors often need to manually measure the size of the gallbladder. However, this method usually relies on the experience and skills of ultrasound physicians, which is not only inefficient, but also difficult to accurately capture key morphological features when faced with complex or irregular gallbladder morphology, resulting in poor measurement accuracy. For this reason, there is an urgent need for a method for automatic recognition and measurement of gallbladder morphology in ultrasound images. Summary of the invention

[0003] In order to overcome the problem of inaccurate manual marking of the longest axis and the shortest axis of the gallbladder when the gallbladder morphology is complex or irregular, the present invention provides a method, system and device for automatic recognition of gallbladder morphology in ultrasound images.

[0004] In a first aspect, in order to solve the above technical problems, the present invention provides a method for automatically identifying gallbladder morphology in an ultrasonic image, comprising:

[0005] The gallbladder region in the gallbladder section is segmented from the ultrasound image using a deep learning segmentation model, and the central skeleton line of the gallbladder is extracted from the gallbladder region using an image skeleton extraction algorithm;

[0006] Obtain two preset points on the central skeleton line, and determine the two end points of the long axis of the gallbladder region according to the positional relationship between the two preset points and all the contour points on the gallbladder region; wherein the two preset points are located at the two end points of the central skeleton line respectively;

[0007] The line segment between the two end points of the major axis is taken as the target major axis;

[0008] The line segment between any two contour points on the gallbladder region is taken as the target line segment, and the target line segment perpendicular to the target long axis and corresponding to the maximum width among the target line segments is taken as the target short axis.

[0009] In a second aspect, the present invention provides an automatic recognition system for gallbladder morphology in ultrasound images, comprising:

[0010] A central skeleton line extraction module is used to segment the gallbladder region in the gallbladder section from the ultrasound image by using a deep learning segmentation model, and to extract the central skeleton line of the gallbladder from the gallbladder region by using an image skeleton extraction algorithm;

[0011] The module for determining the two end points of the long axis is used to obtain two preset points on the central skeleton line, and determine the two end points of the long axis of the gallbladder region according to the positional relationship between the two preset points and all the contour points on the gallbladder region; wherein the two preset points are located at the two end points of the central skeleton line respectively;

[0012] A target long axis acquisition module is used to take the line segment between the two end points of the long axis as the target long axis;

[0013] The target short axis acquisition module is used to take the line segment between any two contour points on the gallbladder area as the target line segment, and take the target line segment perpendicular to the target long axis and corresponding to the maximum width among the target line segments as the target short axis.

[0014] In a third aspect, the present invention provides a computing device comprising a memory, a processor, and a program stored in the memory and running on the processor, wherein when the processor executes the program, the steps of the method for automatic recognition of gallbladder morphology in ultrasound images as described above are implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having instructions stored therein. When the instructions are executed on a terminal device, the terminal device executes the steps of the above-mentioned method for automatic recognition of gallbladder morphology in an ultrasound image.

[0016] The beneficial effects of the present invention are: by segmenting the gallbladder area in the gallbladder section from the ultrasound image, the central skeleton line can be extracted from the gallbladder area, and the target long axis is determined according to the positional relationship between two preset points on the central skeleton line and all contour points on the gallbladder area, and finally the target short axis is determined by the target long axis. In this application, by adding image segmentation technology (deep learning segmentation model) and target extraction algorithm (image skeleton extraction algorithm), the morphology of the gallbladder can be automatically identified, and the target long axis and target short axis of the gallbladder can be automatically identified according to the morphology of the gallbladder in the ultrasound image, so that when the gallbladder morphology is complex or irregular, the accuracy of identifying the longest axis and the shortest axis of the gallbladder is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0018] Figure 1 A schematic diagram of a flow chart of a method for automatically identifying gallbladder morphology in an ultrasonic image according to an embodiment of the present invention;

[0019] Figure 2 It is a schematic diagram of the gallbladder section and the corresponding gallbladder area when there is no inflection point;

[0020] Figure 3 It is a schematic diagram of the gallbladder area and the central skeleton line when there is no inflection point;

[0021] Figure 4 It is a schematic diagram of the preset points and neighboring points when there is no inflection point;

[0022] Figure 5 It is a schematic diagram of the two end points of the long axis when there is no inflection point;

[0023] Figure 6 It is a schematic diagram of the long axis when there is no inflection point;

[0024] Figure 7 It is a schematic diagram of the gallbladder section and the corresponding gallbladder area when there is an inflection point;

[0025] Figure 8 It is a schematic diagram of the gallbladder area and the central skeleton line when there is an inflection point;

[0026] Fig. 9 It is a schematic diagram of the preset points and neighboring points when there is an inflection point;

[0027] Fig.10 It is a schematic diagram of the two end points of the long axis when there is an inflection point;

[0028] Fig.11 It is a schematic diagram of the long axis when there is an inflection point;

[0029] Fig.12 It is a schematic diagram of the major axis and the minor axis when there is no inflection point;

[0030] Fig.13 It is a schematic diagram of the major axis and the minor axis when there is an inflection point;

[0031] Fig.14 The present invention is a schematic diagram of the structure of an automatic gallbladder morphology recognition system in an ultrasonic image according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following examples are provided to further explain and supplement the present invention and do not constitute any limitation to the present invention.

[0033] The following describes a method, system and device for automatically identifying gallbladder morphology in ultrasound images according to an embodiment of the present invention in conjunction with the accompanying drawings.

[0034] like Figure 1 As shown, the present invention provides a method for automatically identifying gallbladder morphology in an ultrasonic image, comprising:

[0035] S1. The gallbladder region in the gallbladder section is segmented from the ultrasound image using a deep learning segmentation model, and the central skeleton line of the gallbladder is extracted from the gallbladder region using an image skeleton extraction algorithm.

[0036] S2. Obtain two preset points on the central skeleton line, and determine the two end points of the long axis of the gallbladder region according to the positional relationship between the two preset points and all the contour points on the gallbladder region; wherein the two preset points are located at the two end points of the central skeleton line respectively.

[0037] S3. The line segment between the two end points of the major axis is taken as the target major axis.

[0038] S4. The line segment between any two contour points on the gallbladder region is taken as the target line segment, and the target line segment perpendicular to the target major axis and corresponding to the maximum width among the target line segments is taken as the target minor axis.

[0039] In this embodiment, the gallbladder region in the gallbladder section is segmented from the ultrasound image, so that the central skeleton line can be extracted from the gallbladder region, and the target long axis is determined according to the positional relationship between two preset points on the central skeleton line and all contour points on the gallbladder region, and finally the target short axis is determined by the target long axis. In this application, by adding image segmentation technology (deep learning segmentation model) and target extraction algorithm (image skeleton extraction algorithm), the morphology of the gallbladder can be automatically identified, and the target long axis and target short axis of the gallbladder can be automatically identified according to the morphology of the gallbladder in the ultrasound image, so that when the gallbladder morphology is complex or irregular, the accuracy of identifying the longest axis and the shortest axis of the gallbladder is improved.

[0040] In this embodiment, the deep learning segmentation model is a common segmentation model, such as the Deeplab neural network, and the image skeleton extraction algorithm is also a common algorithm, such as the Zhang-Suen image skeleton extraction algorithm, so they are not described in detail.

[0041] like Figure 2 As shown, Figure 2 The first column of images in the figure is the acquired ultrasound image of the gallbladder. The gallbladder region in the gallbladder section is segmented from the ultrasound image using a deep learning segmentation model to obtain a gallbladder mask, as shown in Figure 2 As shown in the second column of images.

[0042] like Figure 3 As shown, Figure 3 The first column of images is the acquired gallbladder area. The central skeleton line is extracted by the image skeleton extraction algorithm, and the following is obtained: Figure 3 The center skeleton line shown in the second column of images can finally be obtained as follows Figure 3 The gallbladder region with the central skeleton line is shown in the third column.

[0043] Optionally, obtaining two preset points on the central skeleton line, and determining the two end points of the long axis of the gallbladder region according to the positional relationship between the two preset points and all contour points on the gallbladder region, includes:

[0044] The nth consecutive neighboring point adjacent to the first preset point of the two preset points and pointing to the second preset point is used as the first target neighboring point, and the nth consecutive neighboring point adjacent to the second preset point and pointing to the first preset point is used as the second target neighboring point;

[0045] The vector pointing from the first preset point to the first target neighborhood point is used as the first vector;

[0046] The vector pointing from the first preset point to any contour point on the gallbladder region is used as the second vector;

[0047] The vector pointing from the second preset point to the second target neighborhood point is used as the third vector;

[0048] The vector pointing from the second preset point to any contour point on the gallbladder region is used as a fourth vector;

[0049] The two end points of the long axis of the gallbladder region are determined according to the positional relationship between the first vector and each second vector, and the positional relationship between the third vector and each fourth vector.

[0050] In this embodiment, n is a positive integer and can be set according to actual conditions. Preferably, in this embodiment, n can be 10.

[0051] like Figure 4 As shown, Figure 4 The first column of images in the figure is the gallbladder region with a central skeleton line. For the sake of convenience, the two preset points are selected from the two end points of the central skeleton line for illustration (the same applies below). Figure 4 The second column of images shows P 01 Point is the first preset point. In addition, {P 01 , P 02 ...P 08 , P 09} as its neighboring point, P 09 is the first target neighborhood point, then P 01 P 09 is the first vector, Figure 4 The third column of images shows P 10 Point is the second preset point. In addition, {P 10 , P 11 ...P 18 , P 19} as its neighboring point, P 19 is the second target neighborhood point, then P 10 P 19 is the third vector.

[0052] In addition, the contour points on the gallbladder region include {B0, B1...B i ...B n}, then P 01B i As the second vector, P 10 B i As the fourth vector, where B i Represents any contour point.

[0053] In this embodiment, two preset points are used to form vectors with the target neighborhood point and the contour point respectively, and the two end points of the long axis can be accurately found through the positional relationship between the vectors. The gallbladder image is clearly visible, and the image recognition method is used to avoid the influence of the complex and irregular shape of the gallbladder, making the identification of the two end points of the long axis more accurate.

[0054] Optionally, determining the two end points of the long axis of the gallbladder region according to the positional relationship between the first vector and each second vector, and the positional relationship between the third vector and each fourth vector, includes:

[0055] Obtaining the angles between the first vector and each second vector to determine a first angle;

[0056] Obtaining the angles between the third vector and each fourth vector to determine the second angle;

[0057] Taking a first angle greater than or equal to a first preset value as a first target angle, and taking contour points on the gallbladder region corresponding to each first target angle as a first target point;

[0058] A second angle whose second angle is greater than or equal to a second preset value is used as a second target angle, and contour points on the gallbladder region corresponding to each second target angle are used as second target points;

[0059] The first target point with the shortest distance to the first preset point among the first target points and the second target point with the shortest distance to the second preset point among the second target points are used as the two end points of the major axis.

[0060] In this embodiment, the first preset value and the second preset value can be set according to actual conditions. Preferably, in this embodiment, the first preset value and the second preset value are both 170°.

[0061] like Figure 5 As shown, Figure 5 The first column of images is the gallbladder area with the central skeleton line. Select the first preset point P 01 and the second preset point P 10 After that, we can get the first vector P 01 P 09 , the second vector P 01 B i , the third vector P 10 P 19 and the fourth vector P 10 B i.

[0062] Among them, the first angle between the first vector and each second vector is compared with 170° to determine the first target angle. Similarly, the second angle between the third vector and each fourth vector is compared with 170° to determine the second target angle. Finally, the distance from each first target point to the first preset point is calculated to obtain the following: Figure 5 The first target point B shown in the second column of the image L1 Similarly, the distance from each second target point to the second preset point is calculated to obtain Figure 5 The second target point B shown in the third column of the image L2 , will soon be B L1 and B L2 As the two end points of the major axis, and if Figure 6 As shown, B L1 and B L2 The line segment between the two is taken as the long axis of the gallbladder.

[0063] Optionally, taking a line segment between two end points of the major axis as a target major axis, further comprising:

[0064] The line segment between the two end points of the major axis is taken as the first initial line segment;

[0065] Obtaining the number of points on the first initial line segment within the gallbladder region, and calculating the ratio between the number of points and the length of the first initial line segment;

[0066] If the ratio is less than the third preset value, a target skeleton point is selected from the central skeleton line; the target skeleton point is an inflection point formed on the central skeleton line due to the bending of the gallbladder morphology;

[0067] Calculate the third angle between the vector of the first endpoint and the vector of the second endpoint of the two endpoints of the long axis respectively pointed by the target skeleton point;

[0068] If the third angle is less than the fourth preset value, the line segment between the target skeleton point and the first endpoint is used as the first major axis, and the line segment between the target skeleton point and the second endpoint is used as the second target major axis; wherein the target major axis includes the first major axis and the second major axis.

[0069] In this embodiment, by setting the inflection point, the morphological characteristics of the gallbladder can be more accurately reflected. Even if the gallbladder has a characteristic of excessive curvature, a measurement inflection point is added at the excessively curved position to ensure that the measurement line is inside the gallbladder as much as possible, thereby reducing the error with the actual value of the long axis of the gallbladder.

[0070] In this embodiment, the third preset value can be set according to actual conditions. Preferably, in this embodiment, the third preset value can be 0.9.

[0071] The fourth preset value in this embodiment can be set according to actual conditions. Preferably, the fourth preset value in this embodiment can be 155°. If the third angle is greater than 155°, it can be considered that although the gallbladder shape is curved, the central skeleton line is almost a straight line, and there is no need to set an inflection point.

[0072] like Figure 7 As shown, Figure 7 The first column of images in the figure is the acquired ultrasound image of the gallbladder. The gallbladder region in the gallbladder section is segmented from the ultrasound image using a deep learning segmentation model to obtain a gallbladder mask, as shown in Figure 7 As shown in the second column of images, it can be seen that the gallbladder is in a curved state.

[0073] like Figure 8 As shown, Figure 8 The first column of images is the acquired gallbladder area. The central skeleton line is extracted by the image skeleton extraction algorithm, and the following is obtained: Figure 8 The central skeleton line with a curved shape shown in the second column of the image can finally be obtained as shown in Figure 8 The gallbladder region with a curved central skeleton line is shown in the third column.

[0074] like Fig. 9 As shown, Fig. 9 The second column of images shows P 01 Point is the first preset point. In addition, {P 01 , P 02 ...P 08 , P 09} as its neighboring point, P 09 is the first target neighborhood point, then P 01 P 09 is the first vector, Fig. 9 The third column of images shows P 10 Point is the second preset point. In addition, {P 10 , P 11 ...P 18 , P 19} as its neighboring point, P 19 is the second target neighborhood point, then P 10 P 19 is the third vector.

[0075] like Fig.10 As shown, Fig.10 The first column of images shows the gallbladder region with a curved central skeleton line. Fig.10 The first target point B is obtained from the second column of images L1 , Fig.10 The third column of images is the second target point B L2 , if at this time, directly change BL1 and B L2 The first initial line segment is obtained by connecting them, and a part of the first initial line segment must be located outside the gallbladder region, so it is necessary to select the target skeleton point from the central skeleton line.

[0076] Optionally, target skeleton points are filtered from the central skeleton line, including:

[0077] Get the distance from each skeleton point on the central skeleton line to the first initial line segment;

[0078] The skeleton point corresponding to the maximum value in the distance is taken as the target skeleton point.

[0079] Connection B L1 and B L2 , get the first initial line segment, traverse all the skeleton points {P0...P i ...P n}, find the distance from each skeleton point to the first initial line segment, and the skeleton point P corresponding to the maximum distance max as the target skeleton point.

[0080] like Fig.11 As shown, calculate the vector P max B L1 and vector P max B L2 If the third angle is less than 155°, then P max As the inflection point, and P max B L1 The line segment between is taken as the first major axis, P max B L2 The line segment between them is taken as the second major axis.

[0081] Optionally, a line segment between any two contour points on the gallbladder region is used as a target line segment, and a target line segment perpendicular to the target major axis and corresponding to the maximum width among the target line segments is used as a target minor axis, further comprising:

[0082] If the target long axis has an inflection point, a neighborhood point adjacent to the target skeleton point on the first long axis is obtained as the third target neighborhood point, and a neighborhood point adjacent to the target skeleton point on the second long axis is obtained as the fourth target neighborhood point;

[0083] Draw a straight line perpendicular to the line segment between the third target neighboring point and the fourth target neighboring point, and divide the gallbladder region into a first region and a second region;

[0084] A line segment between any two contour points on the first region is used as a second initial line segment, and a line segment between any two contour points on the second region is used as a third initial line segment;

[0085] The second initial line segment that is perpendicular to the first major axis and corresponds to the maximum width among the second initial line segments is used as the first minor axis, and the third initial line segment that is perpendicular to the second major axis and corresponds to the maximum width among the third initial line segments is used as the second minor axis;

[0086] The lengths of the first minor axis and the second minor axis are compared, and the minor axis with the larger length is taken as the target minor axis.

[0087] like Fig.12 As shown, Fig.12 The first column of images in the figure shows the gallbladder without adding inflection points. The long axis of the gallbladder in this form has no inflection points set, and due to the morphological characteristics of the gallbladder, there is only one short axis with the largest width from top to bottom. Therefore, the corresponding short axis can be directly obtained from the long axis. Fig.12 As shown in the second column of images, finally, we can get Fig.12 The third column of images shows the major and minor axes of this pattern.

[0088] like Fig.13 As shown, Fig.13 The first column of images in the figure is a curved gallbladder region. Since the inflection point is set, the gallbladder region is divided into two regions, the upper and lower regions, by a straight line a perpendicular to the line segment between the third target neighborhood point and the fourth target neighborhood point. The corresponding short axis can be obtained in both regions, as shown in Fig.13 The line segment b shown in the second column of the image is the first minor axis, and the line segment c is the second minor axis. It can be seen that the length of the first minor axis is significantly greater than the second minor axis. Therefore, the second minor axis is taken as the target minor axis, and we get Fig.13 The long and short axes of the gallbladder in this morphology are shown in the third column of images.

[0089] Optionally, the two preset points are two end points of the central skeleton line.

[0090] In this embodiment, the two end points of the central skeleton line are selected as preset points, which can reduce the amount of calculation and improve the efficiency of obtaining the long axis and short axis of the gallbladder.

[0091] like Fig.14 As shown, the present invention provides an automatic recognition system for gallbladder morphology in ultrasound images, comprising:

[0092] A central skeleton line extraction module is used to segment the gallbladder region in the gallbladder section from the ultrasound image by using a deep learning segmentation model, and to extract the central skeleton line of the gallbladder from the gallbladder region by using an image skeleton extraction algorithm;

[0093] The module for determining the two end points of the long axis is used to obtain two preset points on the central skeleton line, and determine the two end points of the long axis of the gallbladder region according to the positional relationship between the two preset points and all the contour points on the gallbladder region; wherein the two preset points are located at the two end points of the central skeleton line respectively;

[0094] A target long axis acquisition module is used to take the line segment between the two end points of the long axis as the target long axis;

[0095] The target short axis acquisition module is used to take the line segment between any two contour points on the gallbladder area as the target line segment, and take the target line segment perpendicular to the target long axis and corresponding to the maximum width among the target line segments as the target short axis.

[0096] Optionally, the module for determining the two end points of the major axis is specifically used for:

[0097] The nth consecutive neighboring point adjacent to the first preset point of the two preset points and pointing to the second preset point is used as the first target neighboring point, and the nth consecutive neighboring point adjacent to the second preset point and pointing to the first preset point is used as the second target neighboring point;

[0098] The vector pointing from the first preset point to the first target neighborhood point is used as the first vector;

[0099] The vector pointing from the first preset point to any contour point on the gallbladder region is used as the second vector;

[0100] The vector pointing from the second preset point to the second target neighborhood point is used as the third vector;

[0101] The vector pointing from the second preset point to any contour point on the gallbladder region is used as a fourth vector;

[0102] The two end points of the long axis of the gallbladder region are determined according to the positional relationship between the first vector and each second vector, and the positional relationship between the third vector and each fourth vector.

[0103] Optionally, the module for determining the two end points of the major axis is specifically used for:

[0104] Obtaining the angles between the first vector and each second vector to determine a first angle;

[0105] Obtaining the angles between the third vector and each fourth vector to determine the second angle;

[0106] Taking a first angle greater than or equal to a first preset value as a first target angle, and taking contour points on the gallbladder region corresponding to each first target angle as a first target point;

[0107] A second angle whose second angle is greater than or equal to a second preset value is used as a second target angle, and contour points on the gallbladder region corresponding to each second target angle are used as second target points;

[0108] The first target point with the shortest distance to the first preset point among the first target points and the second target point with the shortest distance to the second preset point among the second target points are used as the two end points of the major axis.

[0109] Optionally, the target long axis acquisition module is further used for:

[0110] The line segment between the two end points of the major axis is taken as the first initial line segment;

[0111] Obtaining the number of points on the first initial line segment within the gallbladder region, and calculating the ratio between the number of points and the length of the first initial line segment;

[0112] If the ratio is less than the third preset value, a target skeleton point is selected from the central skeleton line; the target skeleton point is an inflection point formed on the central skeleton line due to the bending of the gallbladder morphology;

[0113] Calculate the third angle between the vector of the first endpoint and the vector of the second endpoint of the two endpoints of the long axis respectively pointed by the target skeleton point;

[0114] If the third angle is less than the fourth preset value, the line segment between the target skeleton point and the first endpoint is used as the first major axis, and the line segment between the target skeleton point and the second endpoint is used as the second target major axis; wherein the target major axis includes the first major axis and the second major axis.

[0115] Optionally, the target long axis acquisition module is specifically used to:

[0116] Get the distance from each skeleton point on the central skeleton line to the first initial line segment;

[0117] The skeleton point corresponding to the maximum value in the distance is taken as the target skeleton point.

[0118] Optionally, the target short axis acquisition module is further used for:

[0119] If the target long axis has an inflection point, a neighborhood point adjacent to the target skeleton point on the first long axis is obtained as the third target neighborhood point, and a neighborhood point adjacent to the target skeleton point on the second long axis is obtained as the fourth target neighborhood point;

[0120] Draw a straight line perpendicular to the line segment between the third target neighboring point and the fourth target neighboring point, and divide the gallbladder region into a first region and a second region;

[0121] A line segment between any two contour points on the first region is used as a second initial line segment, and a line segment between any two contour points on the second region is used as a third initial line segment;

[0122] The second initial line segment that is perpendicular to the first major axis and corresponds to the maximum width among the second initial line segments is used as the first minor axis, and the third initial line segment that is perpendicular to the second major axis and corresponds to the maximum width among the third initial line segments is used as the second minor axis;

[0123] The lengths of the first minor axis and the second minor axis are compared, and the minor axis with the larger length is taken as the target minor axis.

[0124] An embodiment of the present invention also provides a computing device, including a memory, a manager, and a program stored in the memory and running on the manager. When the manager executes the program, some or all steps of the above-mentioned method for automatic recognition of gallbladder morphology in ultrasound images are implemented.

[0125] The computing device may be a computer, and correspondingly, its program is computer software. The parameters and steps in the computing device of the present invention mentioned above may refer to the parameters and steps in an embodiment of an automatic recognition method for gallbladder morphology in ultrasound images, which will not be described in detail here.

[0126] Those skilled in the art know that the present invention can be implemented as a system, method or computer program product. Therefore, the present disclosure can be specifically implemented in the following forms, namely: it can be complete hardware, it can be complete software (including firmware, resident software, microcode, etc.), or it can be a combination of hardware and software, generally referred to as "circuit", "module" or "system" herein. In addition, in some embodiments, the present invention can also be implemented in the form of a computer program product in one or more computer-readable media, and the computer-readable medium contains computer-readable program code. Computer-readable storage media can be, for example, but not limited to - electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination of the above.

[0127] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0128] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A method for automatically identifying gallbladder morphology in ultrasonic images, characterized in that: include: Segmenting the gallbladder region in the gallbladder section from the ultrasound image using a deep learning segmentation model, and extracting the central skeleton line of the gallbladder from the gallbladder region using an image skeleton extraction algorithm; Obtaining two preset points on the central skeleton line, and determining the two end points of the long axis of the gallbladder region according to the positional relationship between the two preset points and all the contour points on the gallbladder region; wherein the two preset points are respectively located at the two end points of the central skeleton line; Taking the line segment between the two end points of the major axis as the target major axis; The line segment between any two contour points on the gallbladder region is taken as the target line segment, and the target line segment perpendicular to the target major axis and corresponding to the maximum width among the target line segments is taken as the target minor axis.

2. The method according to claim 1, characterized in that Acquiring two preset points on the central skeleton line, and determining the two end points of the long axis of the gallbladder region according to the positional relationship between the two preset points and all contour points on the gallbladder region, including: The nth consecutive neighboring point adjacent to the first preset point of the two preset points and pointing to the second preset point is used as the first target neighboring point, and the nth consecutive neighboring point adjacent to the second preset point and pointing to the first preset point is used as the second target neighboring point; Taking the vector pointing from the first preset point to the first target neighborhood point as the first vector; The vector pointing from the first preset point to any contour point on the gallbladder region is used as the second vector; Taking the vector pointing from the second preset point to the second target neighborhood point as the third vector; The vector pointing from the second preset point to any contour point on the gallbladder region is used as a fourth vector; The two end points of the long axis of the gallbladder region are determined according to the positional relationship between the first vector and each of the second vectors, and the positional relationship between the third vector and each of the fourth vectors.

3. The method according to claim 2, characterized in that Determining the two end points of the long axis of the gallbladder region according to the positional relationship between the first vector and each of the second vectors, and the positional relationship between the third vector and each of the fourth vectors, comprises: Obtaining the angle between the first vector and each of the second vectors to determine a first angle; Obtaining the angle between the third vector and each of the fourth vectors to determine a second angle; Taking a first angle where the first angle is greater than or equal to a first preset value as a first target angle, and taking contour points on the gallbladder region corresponding to each first target angle as a first target point; A second angle where the second angle is greater than or equal to a second preset value is used as a second target angle, and contour points on the gallbladder region corresponding to each second target angle are used as second target points; The first target point with the shortest distance to the first preset point among the first target points and the second target point with the shortest distance to the second preset point among the second target points are used as the two end points of the major axis.

4. The method according to claim 1, characterized in that: The line segment between the two end points of the major axis is used as the target major axis, and further includes: Taking the line segment between the two end points of the major axis as the first initial line segment; Obtaining the number of points on the first initial line segment within the gallbladder region, and calculating the ratio between the number of points and the length of the first initial line segment; If the ratio is less than a third preset value, a target skeleton point is selected from the central skeleton line; the target skeleton point is an inflection point formed on the central skeleton line due to the bending of the gallbladder morphology; Calculate the third angle between the vector of the first endpoint and the vector of the second endpoint of the two endpoints of the long axis respectively pointed by the target skeleton point; If the third angle is less than a fourth preset value, the line segment between the target skeleton point and the first endpoint is used as the first major axis, and the line segment between the target skeleton point and the second endpoint is used as the second target major axis; wherein the target major axis includes the first major axis and the second major axis.

5. The method according to claim 4, characterized in that Selecting target skeleton points from the central skeleton line includes: Obtaining the distance from each skeleton point on the central skeleton line to the first initial line segment; The skeleton point corresponding to the maximum value of the distance is taken as the target skeleton point.

6. The method according to claim 4, characterized in that The line segment between any two contour points on the gallbladder region is used as a target line segment, and the target line segment perpendicular to the target major axis and corresponding to the maximum width among the target line segments is used as the target minor axis, further comprising: If the target long axis has an inflection point, a neighborhood point adjacent to the target skeleton point on the first long axis is obtained as a third target neighborhood point, and a neighborhood point adjacent to the target skeleton point on the second long axis is obtained as a fourth target neighborhood point; Draw a straight line perpendicular to the line segment between the third target neighborhood point and the fourth target neighborhood point, and divide the gallbladder region into a first region and a second region; A line segment between any two contour points on the first region is used as a second initial line segment, and a line segment between any two contour points on the second region is used as a third initial line segment; The second initial line segment that is perpendicular to the first major axis and corresponds to the maximum width among the second initial line segments is used as the first minor axis, and the third initial line segment that is perpendicular to the second major axis and corresponds to the maximum width among the third initial line segments is used as the second minor axis; The lengths of the first minor axis and the second minor axis are compared, and the minor axis with the larger length is taken as the target minor axis.

7. The method according to any one of claims 1 to 6, characterized in that: The two preset points are respectively the two end points of the central skeleton line.

8. An automatic recognition system for gallbladder morphology in ultrasound images, characterized in that: include: A central skeleton line extraction module, used to segment the gallbladder region in the gallbladder section from the ultrasound image by using a deep learning segmentation model, and to extract the central skeleton line of the gallbladder from the gallbladder region by using an image skeleton extraction algorithm; A module for determining two end points of the long axis, used to obtain two preset points on the central skeleton line, and determine the two end points of the long axis of the gallbladder region according to the positional relationship between the two preset points and all the contour points on the gallbladder region; wherein the two preset points are respectively located at the two end points of the central skeleton line; A target long axis acquisition module, used for taking the line segment between the two end points of the long axis as the target long axis; The target short axis acquisition module is used to take the line segment between any two contour points on the gallbladder area as the target line segment, and take the target line segment perpendicular to the target long axis and corresponding to the maximum width among the target line segments as the target short axis.

9. A computing device comprising a memory, a processor, and a program stored in the memory and running on the processor, characterized in that: When the processor executes the program, the steps of a method for automatically identifying gallbladder morphology in ultrasonic images as claimed in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the steps of a method for automatically identifying gallbladder morphology in an ultrasonic image as claimed in any one of claims 1 to 7.

Images