WO2020154904A1

WO2020154904A1 - Method for automatically measuring blood vessel diameter in ultrasound image

Info

Publication number: WO2020154904A1
Application number: PCT/CN2019/073736
Authority: WO
Inventors: 李聪; 王兴红
Original assignee: 深圳市科曼医疗设备有限公司
Priority date: 2019-01-29
Filing date: 2019-01-29
Publication date: 2020-08-06

Abstract

A method for automatically measuring a blood vessel diameter in an ultrasound image, comprising: acquiring a preprocessed ultrasound image (S100); obtaining a multi-threshold segmented ultrasound image according to the acquired pre-processed ultrasound image (S110); and automatically measuring a blood vessel diameter size via ellipse fitting according to the obtained multi-threshold segmented ultrasound image (S120). After enhancing an ultrasound image, a large number of noise are visible all over the ultrasound image. Noise reduction is performed on the enhanced image, and multi-threshold image segmentation is performed on the noise-reduced image, then the ultrasound image is divided into four different regions according to pixel grayscale values. Since the shape of a blood vessel is generally circular, whereas the shape of a diseased blood vessel is generally elliptical, when the segmented blood vessel regions are subjected to elliptic curve fitting, the algorithm does not require any manual intervention and achieves automatic measurement of the blood vessel diameter in the ultrasound image, thereby providing an important clinical auxiliary diagnostic technique for PICC or CVC surgery.

Description

Automatic measurement method of blood vessel diameter in ultrasound image

Technical field

The invention relates to the technical field of medical image analysis, in particular to a method for automatically measuring blood vessel diameters in ultrasound images.

Background technique

PICC (central venous catheterization via peripheral venous puncture) or CVC (central venous catheter), as a common clinical infusion tool, is a kind of infusion that can be punctured through the skin and stayed in the venous cavities (superior vena cava, inferior vena cava, brachiocephalic , Internal jugular vein, subclavian vein, iliac vein, femoral vein) special silicone tube for long-term infusion. During the operation, if the diameter of the catheter is too large or too small, it will bring adverse consequences to the operation.

In order to be able to select a suitable catheter for intubation according to the diameter of the blood vessel, it is necessary to use an ultrasound device to perform an imaging display of the blood vessel area at the site to be punctured, and to select a suitable cannula for surgery based on the displayed blood vessel diameter information. Traditional measurement methods require experienced doctors to manually draw the blood vessel region of the ultrasound image, which is greatly affected by subjective factors, time-consuming drawing and increases the difficulty of recognition. Ultrasound has become an important tool for clinical diagnosis due to its real-time, non-destructive, and low-cost characteristics. Therefore, ultrasound imaging is more and more widely used in PICC or CVC puncture operations.

At present, the measurement of blood vessel diameter is realized by segmentation and morphological processing based on Otsu. This method requires detection based on HAAR features (ie linear features, edge features, point features, diagonal features) and Adaboost (Adaptive boosting) classifiers. Methods After marking the blood vessel area to be segmented with a square frame, the square center point is used as the blood vessel center point, and the binarized image is obtained through Otsu threshold segmentation, combined with the morphological processing method to obtain the blood vessel boundary, and finally the distance between the center point and the boundary is directly measured , As the radius information of the last blood vessel. This method does not consider the influence of noise on the edges of blood vessels, and the selection of the center point is random, and the calculation accuracy is low. The blood vessel area appears as low echo in the image, and the color is dark, making it difficult to distinguish the blood vessel area from the background area, so the ultrasound image must be enhanced. If the enhancement is too high, the necessary blood vessel information will be lost; if the enhancement is too weak, the blood vessel area cannot be distinguished from the background area.

Therefore, how to automatically measure the diameter of blood vessels in ultrasound images has become an urgent technical problem to be solved.

Summary of the invention

The technical problem to be solved by the present invention is how to automatically measure the blood vessel diameter of the ultrasound image.

To this end, according to the first aspect, an embodiment of the present invention discloses an automatic measurement method of ultrasound image blood vessel diameter, including: acquiring a pre-processed ultrasound image; obtaining a multi-threshold segmentation ultrasound image according to the acquired pre-processed ultrasound image; The ultrasound image is segmented according to the obtained multi-threshold value, and the blood vessel diameter is automatically measured through ellipse fitting.

Optionally, the obtaining the preprocessed ultrasound image includes: performing a fractional differential enhancement calculation on the ultrasound image; performing a denoising operation on the enhanced ultrasound image to obtain a denoised ultrasound image.

Optionally, the performing fractional differential enhancement calculation on the ultrasound image includes: giving the corresponding differential enhancement order v, obtaining the ultrasound image x and y axis gradient values; calculating the ultrasound image x and y axis gradient value average, Obtain enhancement factors for image enhancement.

Optionally, the denoising operation on the enhanced ultrasound image to obtain the denoised ultrasound image includes: based on the differential enhancement order v, giving a diffusion threshold k, and performing anisotropic diffusion filtering on the ultrasound image.

Optionally, the obtaining the multi-threshold segmented ultrasound image according to the pre-processed ultrasound image includes: using a particle swarm optimization algorithm to divide the denoised image according to the gray value of the pixel point, and divide the ultrasound image into four A region with different gray values; binarize the segmented image, set the region with the smallest gray value to 0, and set the rest to 1; use the hole filling method to obtain all connected regions in the binary ultrasound image; calculate The area of each connected area retains the largest area of the ultrasound image; through the edge detection method, the edge of the largest area of the ultrasound image is obtained and displayed on the original ultrasound image.

Optionally, the segmentation of the ultrasound image according to the obtained multi-threshold value, and the automatic measurement of the diameter of the blood vessel through ellipse fitting includes: performing ellipse fitting on the edge points of the segmentation target by the least square method; and displaying the fitting result in the original ultrasound On the image, and automatically calculate the size of the blood vessel diameter according to the fitting result.

According to a second aspect, an embodiment of the present invention discloses an automatic measurement device for blood vessel diameter in ultrasound images, which is characterized in that it comprises: an image preprocessing module for acquiring preprocessed ultrasound images; an image segmentation module for The preprocessed ultrasound image is obtained to obtain a multi-threshold segmentation ultrasound image; the diameter measurement module is used to segment the ultrasound image according to the obtained multi-threshold value, and automatically measure the diameter of the blood vessel through ellipse fitting.

Optionally, the image preprocessing module includes: an image enhancement unit for performing fractional differential enhancement calculation on the ultrasound image; an image denoising unit for performing a denoising operation on the enhanced ultrasound image to obtain denoising Ultrasound image.

According to a third aspect, an embodiment of the present invention discloses a computer device, including a processor, configured to execute a computer program stored in a memory to implement the method for automatically measuring blood vessel diameter in an ultrasound image according to any one of the above-mentioned first aspects.

According to a fourth aspect, an embodiment of the present invention discloses a computer-readable storage medium on which a computer program is stored, and the processor is configured to execute the computer program stored in the storage medium to implement the ultrasound image blood vessel diameter of any one of the above-mentioned first aspects. Automatic measurement method.

The present invention has the following beneficial effects: by performing an enhancement operation on the ultrasound image, since the ultrasound image is filled with a large number of noise particles, the enhanced image is noise smoothed, and then the smoothed image is segmented with multiple thresholds, and the ultrasound image Divided into four different areas according to the pixel gray value. Since the shape of blood vessels is generally circular, and sick blood vessels are generally elliptical, the segmented blood vessel area is fitted with elliptic curve. The entire algorithm does not require manual intervention and realizes the ultrasound image of blood vessels. The automatic measurement of diameter provides an important clinical auxiliary diagnostic technique for PICC or CVC surgery.

Description of the drawings

In order to explain the specific embodiments of the present invention or the technical solutions in the prior art more clearly, the following will briefly introduce the drawings that need to be used in the specific embodiments or the description of the prior art. Obviously, the appendix in the following description The drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained from these drawings without creative work.

FIG. 1 is a schematic flowchart of an automatic method for measuring blood vessel diameter in ultrasound images disclosed in this embodiment;

2 is a schematic structural diagram of an automatic measurement device for blood vessel diameter in ultrasound images disclosed in this embodiment;

3 is a schematic diagram of the steps of a method for automatically measuring blood vessel diameter in ultrasound images disclosed in this embodiment;

Fig. 4 is an ultrasound image comparison effect diagram of a method for automatically measuring blood vessel diameter in an ultrasound image disclosed in this embodiment; Fig. 4a is an original ultrasound image; Fig. 4b is an ultrasound blood vessel labeling image;

FIG. 5 is a fractional differential enhancement template of a method for automatically measuring blood vessel diameter in ultrasound images disclosed in this embodiment;

FIG. 6 is an anisotropic diffusion filter template based on fractional differentiation of an automatic method for measuring blood vessel diameter in ultrasound images disclosed in this embodiment;

Fig. 7 is an ultrasound image preprocessing result diagram of an ultrasonic image blood vessel diameter automatic measurement method disclosed in this embodiment; Fig. 7a is the original ultrasound image; Fig. 7b is the enhanced image; Fig. 7c is the filtered image;

Fig. 8 is an image segmentation process diagram of a method for automatically measuring blood vessel diameter in ultrasound images disclosed in this embodiment; Fig. 8a is a segmentation threshold segmented image; Fig. 8b is a binarized image; Fig. 8c is a connected region image; Fig. 8d is an acquisition The edge image of the largest connected area; Figure 8e is the image of the blood vessel area after segmentation.

Reference signs: steps S100-S130; 210, image preprocessing module; 211, image enhancement unit; 212, image denoising unit; 220, image segmentation module; 230, diameter measurement module.

detailed description

In order to make the objectives, technical solutions, and advantages of the present invention clearer, the following further describes the present invention in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, but not to limit the present invention.

The embodiment of the present invention discloses a method for automatically measuring blood vessel diameter in ultrasound images, as shown in Fig. 1 and Fig. 3, including:

Step S110, acquiring a preprocessed ultrasound image;

Step S120: Obtain a multi-threshold segmented ultrasound image according to the acquired preprocessed ultrasound image;

In step S130, the ultrasound image is segmented according to the obtained multi-threshold value, and the blood vessel diameter is automatically measured through ellipse fitting.

As shown in FIG. 4, FIG. 4 is an ultrasound image comparison effect diagram of an ultrasound image blood vessel diameter automatic measurement method disclosed in this embodiment. FIG. 4a is an original ultrasound image, and FIG. 4b is an ultrasound blood vessel mark image.

Compared with the prior art, the solution disclosed in the embodiments of the present invention performs an enhancement operation on the ultrasound image. Since the ultrasound image is filled with a large number of noise particles, the enhanced image is noise smoothed, and then the smoothed image is multi-threshold image Segmentation, and divide the ultrasound image into four different regions according to the pixel gray value. Since the shape of blood vessels is generally circular, the diseased blood vessels are generally elliptical, and the elliptic curve fitting is performed on the segmented blood vessel area. The entire algorithm does not require manual intervention , To realize the automatic measurement of the diameter of the blood vessel in the ultrasound image, thereby providing an important clinical auxiliary diagnostic technology for PICC or CVC surgery.

In a specific embodiment, step S110 may specifically include:

Step S111, performing fractional differential enhancement calculation on the ultrasound image;

Step S112: Perform a denoising operation on the enhanced ultrasound image to obtain a denoised ultrasound image.

In a specific embodiment, step S111 may specifically include:

Given the corresponding differential enhancement order v, obtain the x and y axis gradient values of the ultrasound image; in this embodiment, the image enhancement adopts the fractional differential algorithm, and the differential expression defined by Grünwld–Letnikov is as follows:

The enhanced template is constructed according to the coefficients on the right side of the equal sign, the center point is set as the center of the mask, and the template is expanded in the positive and negative directions and diagonal directions of the x and y axes to make the template have rotation invariance. As shown in Fig. 5, Fig. 5 is a fractional differential enhancement template of a method for automatically measuring blood vessel diameter in ultrasound images disclosed in this embodiment.

Calculate the mean value of the gradient in the x and y axis of the ultrasound image, and obtain the enhancement factor for image enhancement. The gradient calculation formula for the x and y axis is as follows:

Get the gradient image:

F=Gx+Gy

Mean gradient:

avg=sum(F(:))/(m*n)

Among them, sum represents the sum function, sum(F(:)) is the sum of the gray values of each pixel of the image F, and m and n are the image size. According to the above calculation formula, the corresponding enhancement factor is as follows:

In a specific embodiment, step S112 may specifically include: based on the differential enhancement order v, giving a diffusion threshold k, and performing anisotropic diffusion filtering on the ultrasound image. In this embodiment, image denoising uses an anisotropic diffusion filtering algorithm based on fractional differential (FAD algorithm). The core idea of the algorithm is to introduce fractional differential theory on the basis of anisotropic diffusion, and pass the diffusion threshold k The mutual cooperation with the differential order v achieves the purpose of image denoising and edge preservation. The mathematical expression of anisotropic diffusion is as follows:

Among them, div is the divergence operator,

Is the gradient of the image,

It is the spread function, used to detect the smooth intensity of the image, λ is usually set to 0.2. The expression of the diffusion function is as follows:

In the formula, k is the diffusion threshold. Combined with the theory of fractional order differentiation, the filter template of the present invention is obtained as shown in FIG. 6.

After image enhancement and image denoising, the blood vessel area can be separated from the background area. FIG. 7 is an ultrasound image preprocessing result diagram of a method for automatically measuring blood vessel diameter in an ultrasound image disclosed in this embodiment. FIG. 7a is an original ultrasound image, FIG. 7b is an enhanced image, and FIG. 7c is a filtered image.

In a specific embodiment, step 120 may specifically include:

Step S121, using a particle swarm optimization algorithm to divide the denoised image according to the gray value of the pixel point, and divide the ultrasound image into four areas with different gray values;

Step S122: Binarize the divided image, set the area with the smallest gray value to 0, and set the rest to 1;

Step S123, using the hole filling method to obtain all connected areas in the binary ultrasound image;

Step S124: Calculate the area of each connected area, and reserve the area with the largest area of the ultrasound image;

In step S125, the edge of the area with the largest area of the ultrasound image is obtained by the edge detection method and displayed on the original ultrasound image.

In this embodiment, the particle swarm optimization algorithm (PSO algorithm) obtains three optimal segmentation thresholds. The PSO algorithm is derived from the study of bird predation behavior, that is, initializing a group of particles in the image and giving the particles an initial velocity and position. Calculate the fitness value of each particle according to the fitness function, where the maximum between-class variance is used as the fitness function to calculate, the expression is as follows:

Where

Is the probability of occurrence of category j,

Is the mean of class j,

Is the mean of all taxa. Calculate the fitness value of each particle through the above formula to obtain the individual optimal value and the global optimal value. Then evolve according to the speed and position formula, the formula is as follows:

In the formula, V _i ^t+1 is the updated particle velocity, and V _i ^t is the current particle velocity. w is the inertia weight coefficient, usually set to [0.8～1.2]. If the value of w is selected too large, the global convergence ability is strong, and the local convergence ability is weak; if the value of w is selected too small, the global convergence ability is weak and the local convergence ability is strong. In order to improve the global convergence ability of the algorithm, set w to 1.2. If it is greater than 1.2, it is easy to fall into a local extreme. c1 and c2 are learning factors, also called acceleration constants. r1 and r2 are random numbers between [0~1].

P _i ^t is the individual extreme value, nP _i ^t is the global extreme value,

Is the current particle position,

Is the updated particle position. This loop iterative calculation until the number of iterations reaches the preset value, the calculation ends and the optimal segmentation threshold is output. Since the blood vessel area returns a weak echo signal, the points in the filtered image that are smaller than the minimum threshold can be set to 1, and the rest are set to 0 to obtain a binary image. Then through the hole filling method, the area connected with the boundary is removed, and the area of each connected area is calculated, and the edge curve of the largest connected area is obtained and superimposed on the original image to finally realize the segmentation of ultrasound blood vessels. As shown in Fig. 8, Fig. 8 is an image segmentation process diagram of an ultrasonic image blood vessel diameter automatic measurement method disclosed in this embodiment; Fig. 8a is a segmentation threshold segmented image; Fig. 8b is a binarized image; Fig. 8c is a connected region Image; Figure 8d is to obtain the edge image of the largest connected area; Figure 8e is the image of the blood vessel area after segmentation.

In a specific embodiment, step S130 may specifically include:

Step S131, ellipse fitting is performed on the edge points of the segmentation target by using the least square method;

Step S132, displaying the fitting result on the original ultrasound image, and automatically calculating the size of the blood vessel diameter according to the fitting result.

In this embodiment, because the ultrasound image is covered by a lot of noise, there are many burrs on the periphery of the final segmented blood vessel. Normal blood vessels in the human body usually turn into a circle, and the diseased blood vessels are elliptical. Therefore, it is necessary to set the edge points after segmentation. For ellipse fitting, the least square method is used here. The diameter of the blood vessel can be calculated by obtaining the coordinate information of the ellipse's centroid, the long and short axis and the ellipse inclination angle according to the fitting results. As shown in Figure 4b, Figure 4b is an ultrasound blood vessel label image, which is the final fitted blood vessel.

As shown in FIG. 2, an embodiment of the present invention discloses an automatic measurement device for the diameter of an ultrasound image blood vessel, which is characterized in that it includes: an image preprocessing module 210 for acquiring preprocessed ultrasound images; an image segmentation module 220 for According to the acquired pre-processed ultrasound image, a multi-threshold segmentation ultrasound image is obtained; the diameter measurement module 230 is configured to segment the ultrasound image according to the obtained multi-threshold value, and automatically measure the diameter of the blood vessel through ellipse fitting.

Optionally, the image preprocessing module 210 includes: an image enhancement unit 211, configured to perform a fractional differential enhancement calculation on the ultrasound image; an image denoising unit 222, configured to perform a denoising operation on the enhanced ultrasound image to obtain Denoised ultrasound image.

In addition, an embodiment of the present invention also provides a computer device, and the processor executes computer instructions to implement the following methods:

Obtain a pre-processed ultrasound image; obtain a multi-threshold segmented ultrasound image according to the obtained pre-processed ultrasound image; obtain a multi-threshold segmented ultrasound image according to the obtained multi-threshold ultrasound image, and automatically measure the diameter of blood vessels through ellipse fitting.

Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments and methods can be implemented by instructing relevant hardware through a computer program. The program can be stored in a computer readable storage medium. At this time, it may include the procedures of the above-mentioned method embodiments. Among them, the storage medium can be a magnetic disk, an optical disc, a read-only memory (ROM) or a random access memory (RAM), etc. The computer processor is used to execute the computer program stored in the storage medium to implement the following methods:

Obviously, the above-mentioned embodiments are merely examples for clear description, and are not intended to limit the implementation. For those of ordinary skill in the art, other changes or changes in different forms can be made on the basis of the above description. It is unnecessary and impossible to list all implementation methods here. The obvious changes or changes derived from this are still within the protection scope of the present invention.

Claims

A method for automatically measuring blood vessel diameter in ultrasound images, which is characterized in that it comprises:

Obtain preprocessed ultrasound images;

Obtaining a multi-threshold segmentation ultrasound image according to the acquired preprocessed ultrasound image;

The ultrasound image is segmented according to the obtained multi-threshold value, and the blood vessel diameter is automatically measured through ellipse fitting.
The method for automatically measuring blood vessel diameter in an ultrasound image according to claim 1, wherein said acquiring the preprocessed ultrasound image comprises:

Perform fractional differential enhancement calculation on ultrasound images;

Perform a denoising operation on the enhanced ultrasound image to obtain a denoised ultrasound image.
The method for automatically measuring blood vessel diameters in ultrasound images according to claim 2, wherein said performing fractional differential enhancement calculation on ultrasound images comprises:

Given the corresponding differential enhancement order v, obtain the x and y axis gradient values of the ultrasound image;

Calculate the mean value of the gradient in the x and y axis of the ultrasound image, and obtain the enhancement factor for image enhancement.
The method for automatically measuring blood vessel diameter in an ultrasound image according to claim 3, wherein the denoising operation on the enhanced ultrasound image, and obtaining the denoised ultrasound image comprises:

Based on the differential enhancement order v, the diffusion threshold k is given to perform anisotropic diffusion filtering on the ultrasound image.
The method for automatically measuring the diameter of blood vessels in an ultrasound image according to claim 1, wherein said obtaining a multi-threshold segmentation ultrasound image according to said obtaining the preprocessed ultrasound image comprises:

Use the particle swarm optimization algorithm to divide the denoised image according to the gray value of the pixel, and divide the ultrasound image into four areas with different gray values;

Binarize the segmented image, set the area with the smallest gray value to 0, and set the rest to 1;

Use the hole filling method to obtain all connected areas in the binary ultrasound image;

Calculate the area of each connected area, and retain the largest area of the ultrasound image;

Through the edge detection method, the edge of the largest area of the ultrasound image is obtained and displayed on the original ultrasound image.
The method for automatically measuring blood vessel diameter in an ultrasound image according to claim 1, wherein said segmenting the ultrasound image according to said multi-threshold value, and realizing automatic measurement of blood vessel diameter through ellipse fitting comprises:

Use the least square method to perform ellipse fitting on the edge points of the segmentation target;

The fitting result is displayed on the original ultrasound image, and the blood vessel diameter is automatically calculated according to the fitting result.
An automatic measurement device for blood vessel diameter in ultrasound images, which is characterized in that it comprises:

Image preprocessing module for obtaining preprocessed ultrasound images;

An image segmentation module, configured to obtain a multi-threshold segmentation ultrasound image according to the acquired preprocessed ultrasound image;

The diameter measurement module is used to segment the ultrasound image according to the obtained multi-threshold value, and automatically measure the diameter of the blood vessel through ellipse fitting.
8. The method for automatically measuring blood vessel diameter in an ultrasound image according to claim 7, wherein the image preprocessing module comprises:

Image enhancement unit for performing fractional differential enhancement calculation on ultrasound images;

The image denoising unit is used to perform denoising operations on the enhanced ultrasound image to obtain a denoised ultrasound image.
A computer device characterized by comprising a processor configured to execute a computer program stored in a memory to implement the method for automatically measuring blood vessel diameter in an ultrasound image according to any one of claims 1-6.
A computer-readable storage medium with a computer program stored thereon, wherein the processor is used to execute the computer program stored in the storage medium to implement the method for automatically measuring blood vessel diameter in ultrasound images according to any one of claims 1-6 .