Method and system for intelligent measurement of vessel diameter in ultrasound examination

Abstract

The application relates to the field of medical image analysis, in particular to a method and system for intelligently measuring the diameter of a blood vessel in ultrasonic examination. The method comprises: a method for intelligently measuring the diameter of a blood vessel in ultrasonic examination, which comprises: identifying an image of at least one target blood vessel site from a preprocessed ultrasonic image; obtaining a boundary contour of a blood vessel lumen of a first site based on the first site image; determining target measurement points of the blood vessel diameter of the first site based on the boundary contour and a preset measurement condition corresponding to the first site, and obtaining an image distance between the target measurement points; and determining a physical distance of the blood vessel diameter of the first site according to the image distance between the target measurement points and combining a first size ratio. The application can automatically determine the optimal measurement points of the blood vessel diameter, obtain more accurate blood vessel diameter values, and further reduce the measurement errors caused by manual intervention.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical image analysis, in particular to a method and system for intelligently measuring blood vessel diameter in ultrasonic examination. BACKGROUND

[0002] In recent years, with the introduction of image processing and computer technology, for blood vessel diameter measurement, from the original projection method, photographic method, ophthalmoscope image measurement method and microscope image measurement method, to the technology combined with screen display and computer image processing, to realize the recognition of blood vessels based on ultrasonic image, to measure the blood vessel diameter.

[0003] However, this technical method needs human intervention to realize the recognition of blood vessels, the determination of the diameter measurement point, and the measurement of the distance between the measurement points, and then the diameter of the blood vessel, and due to the complexity of the blood vessel distribution of each part of the human body and the non-uniformity of the blood vessel diameter, the determination of the blood vessel diameter measurement point after judging the part in the image processing is mixed with human subjective consciousness judgment, and the best diameter measurement point cannot be obtained, resulting in a large error in diameter measurement. SUMMARY

[0004] In order to realize the automatic determination of the best blood vessel diameter measurement point and obtain more accurate blood vessel diameter value, and further reduce the measurement error caused by manual intervention, the present application provides a method and system for intelligently measuring blood vessel diameter in ultrasonic examination.

[0005] In a first aspect, the present application provides a method for intelligently measuring blood vessel diameter in ultrasonic examination, which adopts the following technical scheme:

[0006] A method for intelligently measuring blood vessel diameter in ultrasonic examination, comprising: identifying at least one target blood vessel part image from a preprocessed ultrasonic image, wherein the preprocessed ultrasonic image comprises at least an image of a first blood vessel, and the at least one target blood vessel part image comprises a first part image of the first blood vessel;

[0007] Based on the first part image, a boundary contour of the blood vessel lumen of the first part is obtained, wherein the boundary contour is a mask contour of the blood vessel lumen;

[0008] Based on the boundary contour and the preset measurement condition corresponding to the first part, a target measurement point of the blood vessel diameter of the first part is determined, and an image distance between the target measurement points is obtained, wherein the blood vessel diameter is the inner diameter of the blood vessel lumen;

[0009] determine a physical distance of the vessel diameter of the first part according to the image distance between the target measurement points, in combination with a first size ratio, wherein the first size ratio is used to represent a ratio of the image distance to the physical distance in the pre-processed ultrasound image.

[0010] By employing the technical solution described above, the image of at least one target blood vessel part is recognized from the pre-processed ultrasound image, and based on the first part image included in the image of at least one target blood vessel part, the boundary profile of the vessel lumen of the first part can be obtained, and the target measurement points of the vessel diameter of the first part are determined based on the preset measurement condition corresponding to the first part. Then, according to the image distance between the target measurement points, the physical distance of the vessel diameter of the first part is determined in combination with the first size ratio. This can realize automatic determination of the best measurement points of the vessel diameter, so as to obtain a more accurate vessel diameter value, and thus the measurement error caused by manual intervention can be reduced.

[0011] Optionally, the determination of the target measurement points of the vessel diameter of the first part based on the boundary profile and the preset measurement condition corresponding to the first part, and the obtaining of the image distance between the target measurement points, specifically include:

[0012] curve fitting is performed on the boundary profile to obtain a fitted curve image;

[0013] measurement points in the curve image are determined based on the preset measurement condition corresponding to the first part, to obtain the target measurement points of the vessel diameter of the first part.

[0014] By employing the technical solution described above, the curve fitting is performed on the boundary profile to obtain a fitted curve image, which can realize clear and smooth highlighting of the boundary profile of the vessel lumen in the form of a curve, and thus more accurate determination of the target measurement points based on the curve can be realized.

[0015] Optionally, the curve fitting on the boundary profile to obtain a fitted curve image specifically includes:

[0016] pixel average values are obtained based on the pixel points on both sides of the edge profile;

[0017] the edge profile formed by the pixel points on one side of the edge profile which are greater than the pixel average values is taken as the upper vessel wall, and the edge profile formed by the pixel points on the other side which are less than the pixel average values is taken as the lower vessel wall;

[0018] the fitted curve is simulated based on the upper vessel wall and the lower vessel wall.

[0019] By adopting the technical scheme, the upper vessel wall and the lower vessel wall of the blood vessel can be quickly distinguished based on the pixel average value of the two side pixel points of the edge contour, so as to accelerate the efficiency of curve fitting and further improve the determination efficiency of the target measurement point.

[0020] Optionally, before the image distance between the target measurement points is obtained, specifically comprising:

[0021] Identifying the position of the scale in the pre-processed ultrasound image;

[0022] Cutting the region of the pre-processed ultrasound image corresponding to the scale position to obtain a target cutting region;

[0023] Matching the target cutting region with each scale line of the scale to obtain the number of pixels existing between the scales in the scale;

[0024] Obtaining the pixel distance based on the number of pixels existing between the scales in the scale and the distance;

[0025] Obtaining the number of pixels existing between the target measurement points, and combining the pixel distance to obtain the image distance between the target measurement points.

[0026] By adopting the technical scheme, the position of the scale is identified, and the scale line of the scale is regionally matched with the pre-processed ultrasound image, so that the number of pixels existing between the scales in the scale can be obtained, and then based on the precision value between the scales in the scale, the pixel distance between the scales can be obtained, and based on the number of pixels existing between the target measurement points, the image distance between the target measurement points can be accurately obtained.

[0027] Optionally, the physical distance of the blood vessel diameter of the first part is determined according to the image distance between the target measurement points and the first size ratio, specifically comprising:

[0028] The pre-processed ultrasound image is scaled down or up based on the size ratio of the scale as the first size ratio;

[0029] The image distance between the target measurement points is multiplied by the first size ratio to obtain the physical distance of the blood vessel diameter of the first part.

[0030] By adopting the technical scheme, the real physical distance of the blood vessel diameter of the first part can be accurately obtained by multiplying the image distance between the target measurement points by the first size ratio.

[0031] Optionally, the image of at least one target blood vessel part is identified from the pre-processed ultrasound image, specifically comprising:

[0032] Obtaining a pre-processed ultrasound image;

[0033] identifying a region where a blood vessel is located in the pre-processed ultrasound image based on a feature site recognition algorithm to obtain at least one blood vessel site;

[0034] cutting out a region corresponding to the at least one blood vessel site from the pre-processed ultrasound image to obtain an image of the at least one target blood vessel site.

[0035] By adopting the above technical solution, the feature information of the blood vessel site in the pre-processed ultrasound image can be recognized by the feature site recognition algorithm, and the at least one site recognized can be processed respectively by cutting out the region corresponding to the at least one blood vessel site from the pre-processed ultrasound image.

[0036] Optionally, before the image of the at least one target blood vessel site is obtained, specifically comprising:

[0037] labeling a blood vessel site name on the image of the at least one blood vessel site, and obtaining a matching confidence of the blood vessel site name label;

[0038] comparing the matching confidence with a preset name matching confidence, and selecting the image of the at least one blood vessel site that meets the preset name confidence as the image of the at least one target blood vessel site.

[0039] By adopting the above technical solution, the blood vessel site that does not meet the preset name matching confidence can be pre-screened by labeling the blood vessel site name and judging the matching confidence of the blood vessel site name label, so as to avoid the error of the final blood vessel diameter detection caused by the inaccurate blood vessel site recognition.

[0040] Optionally, the first site image is used to obtain a boundary contour of a blood vessel lumen of the first site, specifically comprising:

[0041] the blood vessel lumen of the first site is recognized based on a blood vessel lumen segmentation algorithm from the first site image;

[0042] a corresponding lumen mask is generated based on the structure of the blood vessel lumen of the first site;

[0043] a boundary contour of the lumen mask is extracted based on a contour extraction algorithm to obtain the boundary contour of the blood vessel lumen of the first site.

[0044] By adopting the technical scheme, the first part blood vessel lumen can be recognized by the first part image blood vessel lumen segmentation algorithm, and the corresponding lumen mask is generated based on the structure of the blood vessel lumen, and the boundary contour of the lumen mask is extracted based on the contour extraction algorithm, so that the boundary contour of the first part blood vessel lumen is obtained, thereby facilitating subsequent processing of the boundary contour to determine the target measurement point of the blood vessel diameter.

[0045] In a second aspect, the application further provides a system for intelligently measuring blood vessel diameter in ultrasonic examination, which adopts the following technical scheme:

[0046] The system for intelligently measuring blood vessel diameter in ultrasonic examination comprises: a part image acquisition module, configured to recognize at least one target blood vessel part image from a preprocessed ultrasonic image, wherein the preprocessed ultrasonic image at least comprises an image of a first blood vessel, and the at least one target blood vessel part image comprises a first part image of the first blood vessel; a boundary contour acquisition module, configured to obtain a boundary contour of a blood vessel lumen of the first part based on the first part image, wherein the boundary contour is a mask contour of the blood vessel lumen; a target measurement point determination module, configured to determine a target measurement point of a blood vessel diameter of the first part based on the boundary contour and a preset measurement condition corresponding to the first part, and obtain an image distance between the target measurement points, wherein the blood vessel diameter is an inner diameter of the blood vessel lumen; and a blood vessel diameter acquisition module, configured to determine a physical distance of the blood vessel diameter of the first part according to the image distance between the target measurement points and in combination with a first size ratio, wherein the first size ratio is used to represent a ratio of an image distance to a physical distance in the preprocessed ultrasonic image.

[0047] By adopting the technical scheme, the at least one target blood vessel part image can be recognized from the preprocessed ultrasonic image by the part image acquisition module, so as to realize recognition of the target blood vessel part, the boundary contour of the blood vessel lumen of the first part can be obtained based on the first part image by the boundary contour acquisition module, the target measurement point of the blood vessel diameter of the first part can be determined based on the boundary contour and the preset measurement condition corresponding to the first part by the target measurement point determination module, and then the image distance between the target measurement points is obtained, and the physical distance of the blood vessel diameter of the first part can be determined according to the image distance between the target measurement points and in combination with the first size ratio by the blood vessel diameter acquisition module.

[0048] Optionally, the target measurement point determination module comprises:

[0049] a curve fitting unit, configured to perform curve fitting on the boundary contour to obtain a fitted curve image;

[0050] A target measurement point determination unit is configured to determine measurement points in the curve image based on the preset measurement condition corresponding to the first part, to obtain the target measurement point of the blood vessel diameter corresponding to the first part.

[0051] By using the above technical solution, the curve fitting unit can realize curve fitting of the boundary profile to obtain a fitted curve image, so as to clearly and smoothly highlight the boundary profile of the blood vessel lumen in the form of a curve. The target measurement point determination unit can more accurately determine the target measurement point based on the curve.

[0052] In summary, the present application has at least one of the following beneficial technical effects:

[0053] 1. By identifying at least one target blood vessel part image from the preprocessed ultrasound image, based on the first part image included in the at least one target blood vessel part image, the boundary profile of the blood vessel lumen of the first part can be obtained, and based on the preset measurement condition corresponding to the first part, the target measurement point of the blood vessel diameter of the first part can be determined. Then, according to the image distance between the target measurement points, the physical distance of the blood vessel diameter of the first part can be determined in combination with the first size ratio, so as to realize automatic determination of the best measurement point of the blood vessel diameter, to obtain a more accurate blood vessel diameter value, and thus the measurement error caused by manual intervention can be reduced.

[0054] 2. By curve fitting the boundary profile, a fitted curve image can be obtained, so as to clearly and smoothly highlight the boundary profile of the blood vessel lumen in the form of a curve, and thus the target measurement point can be more accurately determined based on the curve.

[0055] 3. By obtaining the average value of the pixel points on both sides of the edge profile, the upper and lower tube walls of the blood vessel can be quickly distinguished, so as to speed up the efficiency of curve fitting, and further improve the efficiency of target measurement point determination. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 is a flowchart of a method for intelligently measuring blood vessel diameter in ultrasonic examination disclosed by the embodiments of the present application;

[0057] Figure 2 is an example schematic diagram of a target measurement point obtained by a method for intelligently measuring blood vessel diameter in ultrasonic examination disclosed by the embodiments of the present application;

[0058] Figure 3 is Figure 1 is an implementation flowchart of the S10 step shown in FIG. 10;

[0059] Figure 4 is Figure 1An implementation flowchart of the S20 step shown in FIG. 2;

[0060] Figure 5 For Figure 1 An implementation flowchart of the S30 step shown in FIG. 3;

[0061] Figure 6 For Figure 5 An implementation flowchart of the S31 step shown in FIG. 4;

[0062] Figure 7 An implementation flowchart of the step of determining the target measurement point in the method for intelligently measuring the vessel diameter in the ultrasonic examination disclosed by another embodiment of the present application;

[0063] Figure 8 For Figure 1 An example flowchart of the step of obtaining the image distance between the target measurement points in the S30 step shown in FIG. 5;

[0064] Figure 9 A module schematic diagram of the system for intelligently measuring the vessel diameter in the ultrasonic examination disclosed by the embodiment of the present application.

[0065] Explanation of reference signs:

[0066] 10, part image acquisition module; 20, boundary contour acquisition module; 30, target measurement point determination module; 40, vessel diameter acquisition module; A, target measurement point of the upper segment of the carotid artery vessel; B, target measurement point of the middle segment of the carotid artery vessel; C, target measurement point of the lower segment of the carotid artery vessel. DETAILED DESCRIPTION

[0067] The present application will be further described in detail below with reference to the accompanying drawings.

[0068] In order to realize automatic determination of the optimal measurement point of the vessel diameter and reduce the measurement error caused by manual intervention, an embodiment of the present application provides a method for intelligently measuring the vessel diameter in the ultrasonic examination. Referring to FIG. 1, Figure 1 The method comprises the following steps:

[0069] S10: identifying at least one image of a target vessel part from a preprocessed ultrasonic image;

[0070] The preprocessed ultrasonic image comprises at least an image of a first vessel, and the image of the at least one target vessel part comprises a first part image of the first vessel.

[0071] In the embodiment, the preprocessed ultrasound image is a carotid artery ultrasound image, and the at least one target blood vessel part identified can be a carotid artery vessel starting segment, a carotid artery vessel middle segment and a carotid artery vessel distal segment. The carotid artery vessel starting segment can be understood as a branched intracranial segment located at the proximal end. The carotid artery vessel middle segment can be understood as a middle segment between the carotid artery vessel starting segment and the carotid artery vessel distal segment, and the middle segment appears as a ball. The carotid artery vessel distal segment is located at the distal end.

[0072] S20: Obtain a boundary contour of a blood vessel lumen of the first part based on the first part image;

[0073] The boundary contour is a mask contour of the blood vessel lumen, and the mask contour can be understood as an upper and lower contour of the blood vessel lumen.

[0074] S30: Determine a target measurement point of a blood vessel diameter of the first part based on the boundary contour and a preset measurement condition corresponding to the first part, and obtain an image distance between the target measurement points;

[0075] The blood vessel diameter can be understood as an inner diameter of the blood vessel lumen, and the preset measurement condition is used for conditional judgment for different parts of the identified carotid artery. In the embodiment, the measurement condition corresponding to different parts can be: when the blood vessel part is the carotid artery vessel starting segment, the measurement point corresponding to the shortest distance inner diameter in the starting segment is taken as the target measurement point; when the blood vessel part is the carotid artery vessel middle segment, the measurement point corresponding to the longest distance inner diameter in the middle segment is taken as the target measurement point; and when the blood vessel part is the carotid artery vessel distal segment, the inner diameter distances of the proximal end and the distal end of the distal segment are compared, and the measurement point corresponding to the shorter inner diameter is taken as the target measurement point. Of course, in the embodiment, the preset measurement condition is not limited, and can be artificially set according to actual conditions.

[0076] S40: Determine a physical distance of the blood vessel diameter of the first part according to the image distance between the target measurement points and in combination with a first size ratio;

[0077] The first size ratio is used to represent the ratio of the image distance to the physical distance in the preprocessed ultrasound image.

[0078] Specifically, in the embodiment, step S40 specifically includes:

[0079] The preprocessed ultrasound image is scaled down or up based on the size ratio of the ruler to obtain the first size ratio;

[0080] The image distance between the target measurement points is multiplied by the first size ratio to obtain the physical distance of the blood vessel diameter of the first part.

[0081] For example, refer to Figure 2 , the ultrasound image of the carotid artery, A represents the target measurement point of the distal segment of the common carotid artery vessel, the distance is 0.339 cm, that is, the common carotid artery vessel distal segment diameter is 0.339 cm, B represents the target measurement point of the bulb segment of the common carotid artery vessel, the distance is 0.565 cm, that is, the common carotid artery vessel bulb segment diameter is 0.565 cm, C represents the target measurement point of the distal segment of the common carotid artery vessel, the distance is 0.517 cm, that is, the common carotid artery vessel distal end diameter is 0.517 cm.

[0082] Refer to Figure 3 In another embodiment, step S10 specifically comprises:

[0083] S11: Obtain a pre-processed ultrasound image; wherein the ultrasound image is a carotid artery image of a patient.

[0084] S12: Identify the region where the blood vessel is located in the pre-processed ultrasound image based on a feature part recognition algorithm, to obtain at least one blood vessel part;

[0085] Wherein, the recognition of the feature part can be realized by a target detection algorithm of a convolution network, and the target detection algorithm may, for example, adopt a detection algorithm such as YOLOv5, R-CNN (Regions-CNN), Fast R-CNN, etc. Of course, in the present embodiment, it is not limited, and the same technical effect can be realized.

[0086] S13: The region corresponding to the at least one blood vessel part is cut out from the pre-processed ultrasound image, to obtain an image of the at least one target blood vessel part.

[0087] Wherein, the region corresponding to the at least one blood vessel part is cut out from the pre-processed ultrasound image, which can be realized by a semantic segmentation method, and the semantic segmentation method may, for example, adopt a method such as PSPNet (Pyramid Scene Parsing Network), Unet network model, and DeeplabV3+ image semantic segmentation. Of course, the segmentation method is not limited herein, and the same technical effect can be realized. In another embodiment, the recognition of the feature part and the cutting operation can also be realized at one time by implementing the segmentation method.

[0088] In addition, in another embodiment, before obtaining the image of the at least one target blood vessel part in step S13, it can specifically comprise:

[0089] Labeling the image of the at least one blood vessel part with a blood vessel part name, and obtaining a matching confidence of the blood vessel part name label;

[0090] The matching confidence of the name is compared with a preset name matching confidence, and an image of the at least one blood vessel part meeting the preset name confidence is selected as an image of the at least one target blood vessel part.

[0091] The name of the blood vessel part is marked, and the matching confidence of the name of the blood vessel part is judged, so that the blood vessel part with a confidence greater than 90%, for example, can be pre-screened, thereby increasing the accuracy of identification of the target blood vessel part.

[0092] Referring to Figure 4 In another embodiment, step S20 specifically includes the following steps:

[0093] S21: The first part image is used to identify the blood vessel lumen of the first part based on a blood vessel lumen segmentation algorithm.

[0094] The blood vessel lumen segmentation algorithm may, for example, use unet++ algorithm, PSPNet (Pyramid Scene Parsing Network), Unet network model, and DeeplabV3+ image semantic segmentation, so as to identify the blood vessel lumen of the first part. Of course, this does not limit the same technical effects.

[0095] S22: A corresponding lumen mask is generated based on the structure of the blood vessel lumen of the first part.

[0096] The lumen mask can be understood as a mask covering the blood vessel lumen.

[0097] S23: The boundary contour of the lumen mask is extracted based on a contour extraction algorithm to obtain the boundary contour of the blood vessel lumen of the first part.

[0098] The contour extraction algorithm may, for example, use Canny operator (multistage edge detection algorithm), Sobel operator (Sobel operator), Laplacian operator (Laplacian Operator), Roberts operator (Roberts operator), and other edge detection methods to extract the boundary contour of the lumen mask. Of course, this does not limit the same technical effects.

[0099] Referring to Figure 5 In another embodiment, step S30 specifically includes the following steps:

[0100] S31: The boundary contour is curve-fitted to obtain a fitted curve image.

[0101] S32: determining a measurement point in the curve image based on the preset measurement condition corresponding to the first part, to obtain the target measurement point corresponding to the first part.

[0102] The curve image of the boundary contour corresponding to the blood vessel lumen is obtained by curve fitting the boundary contour through step S31, so that the boundary contour of the blood vessel lumen can be clearly and smoothly highlighted in the form of a curve, and the target measurement point is more accurately determined based on the curve in step S32.

[0103] Specifically, referring to Figure 6 , step S31 can further specifically include:

[0104] S311: obtaining a pixel average value based on the two side pixel points of the edge contour;

[0105] In this embodiment, the pixel average value can be obtained by establishing a pixel coordinate axis based on the near and far ends of the preprocessed ultrasound image, and the average value of the two side pixel points is calculated based on the coordinate positions.

[0106] S312: forming the edge contour of the side pixel point greater than the pixel average value as the front vessel wall of the blood vessel, and forming the edge contour of the side pixel point less than the pixel average value as the rear vessel wall of the blood vessel; wherein the front vessel wall and the rear vessel wall of the blood vessel are determined based on the emission surface of the ultrasound probe, and the vessel wall close to the emission surface of the ultrasound probe is regarded as the front vessel wall, and the other side wall far away from the emission surface of the ultrasound probe is regarded as the rear vessel wall.

[0107] S313: simulating the fitting curve based on the upper vessel wall and the lower vessel wall of the blood vessel.

[0108] In the above steps S311-S313, the front vessel wall and the rear vessel wall of the blood vessel can be quickly distinguished based on the pixel average value, so as to accelerate the efficiency of curve fitting, and further improve the efficiency of determining the target measurement point.

[0109] For example, referring to Figure 7, the first column (1) on the left side of the figure sequentially from top to bottom is the image of the middle segment of the common carotid artery, the initial segment of the common carotid artery and the proximal segment of the common carotid artery, the boundary contour of the lumen is extracted by the contour extraction algorithm, as shown in the second column (2), then the front and rear walls of the boundary contour are extracted, as shown in the third column (3), then the curve fitting of the front and rear walls of the lumen is performed, as shown in the fourth column (4), and the average curve between the curves between the upper and lower walls in the curve fitting image is used to distinguish the front and rear walls of the lumen, and finally the curve fitting image is combined with the preset measurement condition of the corresponding part to obtain the target measurement point image of each part corresponding to the image in the first column (1) on the left side, as shown in the fifth column (5).

[0110] Referring to Figure 8 In another embodiment, before obtaining the image distance between the target measurement points in step S30, it further specifically comprises:

[0111] S301: identifying the position of the ruler in the preprocessed ultrasound image;

[0112] S302: cutting the region of the preprocessed ultrasound image corresponding to the ruler position to obtain a target cutting region.

[0113] S303: matching the target cutting region with each scale line of the ruler to obtain the number of pixels existing between the scales in the ruler.

[0114] S304: obtaining the pixel distance based on the number of pixels existing between the scales in the ruler and the distance.

[0115] S305: obtaining the number of pixels existing between the target measurement points, and combining the pixel distance to obtain the image distance between the target measurement points.

[0116] Wherein, the position of the ruler relative to the preprocessed ultrasound image can be above, below, left side, right side, etc., and the identification of the position of the ruler can be performed by a trained model of prior knowledge mechanism, in this embodiment, the ruler is a ruler with scale lines for size reference, the scale lines include long scale lines and short scale lines, the distance between the long scale lines is, for example, 1 cm, and the distance between the short scale lines is, for example, 0.5 cm.

[0117] For example, the distance between two short scale lines is d1, the number of pixels is d2, the number of pixels between the target measurement points is d3, and the image distance of the target measurement points is d4, then d4 = d1 * d2 / d3.

[0118] To sum up, the method for intelligently measuring a blood vessel diameter in ultrasonic examination disclosed in the application can automatically determine the optimal measurement point of the blood vessel diameter to obtain a more accurate blood vessel diameter value, and thus can reduce the measurement error caused by manual intervention. The boundary profile of the blood vessel lumen can be clearly and smoothly highlighted in the form of a curve, and the target measurement point can be more accurately determined based on the curve. The average pixel value obtained based on the pixel points on both sides of the boundary profile can quickly distinguish the upper and lower vessel walls of the blood vessel, thereby speeding up the efficiency of the curve fitting and further improving the efficiency of the target measurement point determination.

[0119] In addition, the labels of the steps in the embodiments are only for convenient description, and do not represent the limitation on the execution sequence of the steps. In actual application, the execution sequence of the steps can be adjusted or performed simultaneously according to the needs, and these adjustments or replacements all belong to the protection scope of the application.

[0120] Another embodiment of the application further provides a system for intelligently measuring a blood vessel diameter in ultrasonic examination. Referring to Figure 9 , the system comprises a part image acquisition module 10, a boundary profile acquisition module 20, a target measurement point determination module 30 and a blood vessel diameter acquisition module 40.

[0121] The part image acquisition module 10 is configured to identify at least one target blood vessel part image from a preprocessed ultrasonic image, wherein the preprocessed ultrasonic image at least comprises an image of a first blood vessel, and the at least one target blood vessel part image comprises a first part image of the first blood vessel. The boundary profile acquisition module 20 is configured to obtain a boundary profile of a blood vessel lumen of the first part based on the first part image, wherein the boundary profile is a two-dimensional planar image of the blood vessel lumen. The target measurement point determination module 30 is configured to determine target measurement points of the blood vessel diameter of the first part based on the boundary profile and a preset measurement condition corresponding to the first part, and obtain an image distance between the target measurement points, wherein the blood vessel diameter is the inner diameter of the blood vessel lumen. The blood vessel diameter acquisition module 40 is configured to determine a physical distance of the blood vessel diameter of the first part according to the image distance between the target measurement points and in combination with a first size ratio, wherein the first size ratio is used to represent the ratio of the image distance to the physical distance in the preprocessed ultrasonic image.

[0122] Further, the target measurement point determination module 30 comprises a curve fitting unit and a target measurement point determination unit. The curve fitting unit is configured to perform curve fitting on the boundary profile to obtain a fitted curve image. The target measurement point determination unit is configured to determine a measurement point in the curve image based on the preset measurement condition corresponding to the first part, to obtain the target measurement of the blood vessel diameter corresponding to the first part.

[0123] It should be noted that the system for intelligently measuring blood vessel diameter in ultrasonic examination disclosed in the embodiment, the method for intelligently measuring blood vessel diameter in ultrasonic examination realized by the system is as described in the foregoing embodiments, and thus will not be described in detail here. Alternatively, each module, unit and other operations or functions in the embodiment can be used to realize the method in the foregoing embodiments.

[0124] The above are preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in the specification (including the abstract and drawings) can be replaced by other equivalent or similar features, unless specifically described. That is, each feature is only an example of a series of equivalent or similar features, unless specifically described.

Claims

1. A method of intelligently measuring vessel diameter in an ultrasound examination, characterized by, The method comprises the following steps: An image of at least one target blood vessel site is identified from a preprocessed ultrasound image, wherein the preprocessed ultrasound image comprises at least an image of a first blood vessel, and the image of the at least one target blood vessel site comprises a first site image of the first blood vessel, and the preprocessed ultrasound image is a carotid ultrasound image; A boundary contour of a blood vessel lumen of the first site is obtained based on the first site image, wherein the boundary contour is a mask contour of the blood vessel lumen; Target measurement points of a blood vessel diameter of the first site are determined based on the boundary contour and a preset measurement condition corresponding to the first site, and an image distance between the target measurement points is obtained, wherein the blood vessel diameter is an inner diameter of the blood vessel lumen, the preset measurement condition is used for conditional judgment for different sites of the identified carotid blood vessel, and the measurement condition corresponding to the different sites is that, when the blood vessel site is a starting segment of a common carotid artery, the target measurement points are taken as measurement points corresponding to the inner diameter of the shortest distance in the starting segment of the blood vessel; when the blood vessel site is a middle segment of the common carotid artery, the target measurement points are taken as measurement points corresponding to the inner diameter of the longest distance in the middle segment of the blood vessel; and when the blood vessel site is a distal segment of the common carotid artery, the inner diameter distances of the proximal end and the distal end of the distal segment of the blood vessel are compared, and the target measurement points are taken as measurement points corresponding to the shorter inner diameter; A physical distance of the blood vessel diameter of the first site is determined according to the image distance between the target measurement points and in combination with a first size ratio, wherein the first size ratio is used to represent the ratio of the image distance to the physical distance in the preprocessed ultrasound image; The method further comprises the following steps: A fitted curve image is obtained by curve fitting the boundary contour; The target measurement points corresponding to the blood vessel diameter of the first site are determined by determining measurement points in the curve image based on the preset measurement condition corresponding to the first site; The method further comprises the following steps: An average pixel value is obtained based on two side pixel points of the boundary contour; An upper blood vessel wall is formed by an edge contour of one side pixel point of the boundary contour which is greater than the average pixel value, and a lower blood vessel wall is formed by an edge contour of one side pixel point of the boundary contour which is less than the average pixel value; The fitted curve is simulated based on the upper blood vessel wall and the lower blood vessel wall.

2. The method of claim 1, wherein, Before the image distance between the target measurement points is obtained, the method further comprises the following steps: A position of a scale in the preprocessed ultrasound image is identified; A target cropped region is obtained by cropping a region of the preprocessed ultrasound image corresponding to the scale position; A number of pixels existing between scales in the scale is obtained by matching the target cropped region with each scale line of the scale; A pixel distance is obtained based on the number of pixels existing between the scales in the scale and a distance. Obtaining the number of pixels existing in the target measurement points, and combining the pixel distance to obtain an image distance between the target measurement points.

3. The method of claim 2, wherein, The image distance between the target measurement points is combined with a first size ratio to determine a physical distance of the first part of the blood vessel diameter, and specifically includes: The pre-processed ultrasound image is scaled down or up based on the size ratio of the ruler as a first size ratio; The image distance between the target measurement points is multiplied by the first size ratio to obtain the physical distance of the first part of the blood vessel diameter.

4. The method according to claim 1, wherein The image of at least one target blood vessel part is identified from the pre-processed ultrasound image, and specifically includes: Obtaining a pre-processed ultrasound image; Identifying the region where the blood vessel is located in the pre-processed ultrasound image based on a feature part identification algorithm to obtain at least one blood vessel part; The region corresponding to the at least one blood vessel part is cut out from the pre-processed ultrasound image to obtain the image of the at least one target blood vessel part.

5. The method of claim 4, wherein, Before obtaining the image of the at least one target blood vessel part, specifically includes: The image of the at least one blood vessel part is marked with a blood vessel part name, and a matching confidence of the blood vessel part name marking is obtained; The matching confidence is compared with a preset name matching confidence, and the image of the at least one blood vessel part that meets the preset name confidence is selected as the image of the at least one target blood vessel part.

6. The method of claim 1, wherein, The first part image is based on the blood vessel lumen segmentation algorithm to identify the blood vessel lumen of the first part, and specifically includes: The first part image is based on the blood vessel lumen segmentation algorithm to identify the blood vessel lumen of the first part; Based on the structure of the blood vessel lumen of the first part, a corresponding lumen mask is generated; Based on the contour extraction algorithm, the boundary contour of the lumen mask is extracted to obtain the boundary contour of the blood vessel lumen of the first part.

7. A system for intelligent measurement of vessel diameter in ultrasound examinations, characterized by For performing the method of any one of claims 1-6, comprising: A part image acquisition module is used to identify at least one target blood vessel part image from a pre-processed ultrasound image, wherein the pre-processed ultrasound image includes at least an image of a first blood vessel, and the image of the at least one target blood vessel part includes a first part image of the first blood vessel; A boundary contour acquisition module is used to obtain a boundary contour of a blood vessel lumen of a first part based on the first part image, wherein the boundary contour is a mask contour of the blood vessel lumen; A target measurement point determination module is used to determine target measurement points of the blood vessel diameter of the first part based on the boundary contour and a preset measurement condition corresponding to the first part, and to obtain an image distance between the target measurement points, wherein the blood vessel diameter is the inner diameter of the blood vessel lumen; A blood vessel diameter acquisition module is used to determine a physical distance of the blood vessel diameter of the first part based on the image distance between the target measurement points and a first size ratio, wherein the first size ratio is used to represent the ratio of the image distance to the physical distance in the pre-processed ultrasound image.

8. The system of claim 7, wherein, The target measurement point determination module includes: a curve fitting unit, configured to perform curve fitting on the boundary profile to obtain a fitted curve image; a target measurement point determination unit, configured to determine a measurement point in the curve image based on the preset measurement condition corresponding to the first part, to obtain the target measurement point corresponding to the blood vessel diameter of the first part.

Images