Methods for determining the endocardial border on echocardiography

By setting the virtual center line on the M-type echocardiography and adjusting the binarization threshold, the endometrium boundary of the heart is accurately determined, which solves the problem of low accuracy of automatic measurement of LVEF in the palm ultrasound machine, and achieves efficient and accurate automatic calculations.

CN114882061BActive Publication Date: 2024-11-12HEALSON TECH CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202110637593.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-06-08
Publication Date
2024-11-12
Estimated Expiration
2041-06-08

AI Technical Summary

Technical Problem

In a palm ultrasound machine, it is difficult to manually measure the left ventricular ejaculation fraction (LVEF), and the prior art cannot accurately and automatically calculate the endometrial boundary of the heart in the M-type echocardiography, resulting in low calculation accuracy.

Method used

By setting the virtual centerline on the M-type echocardiography, segmenting the image and adjusting the binarization threshold, combining pixel grayscale analysis, accurately determining the boundaries of the left ventricular septum and the posterior left ventricular wall, eliminating the mitral valve image Influence, realize automatic calculation of LVEF.

Benefits of technology

The accuracy and efficiency of LVEF automatic measurement in M-type echocardiography is improved. The automatic measurement time is 0.6 seconds, and the accuracy rate is 95.6%, exceeding the accuracy of manual calculation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114882061B_ABST
    Figure CN114882061B_ABST
Patent Text Reader

Abstract

The present invention provides a method for determining the endocardial border on an echocardiogram. The method first reads an original M-type echocardiogram, then sets a virtual center line on the echocardiogram, and divides the method into an upper and lower halves from the center line, which are an upper echocardiogram having a mitral valve image and a left ventricular septum endocardium image, respectively; and obtains the left ventricular septum border by processing the upper echocardiogram. And the lower echocardiogram having the endocardial border of the left ventricular posterior wall. The lower echocardiogram is processed to obtain the endocardial border of the left ventricular posterior wall. The method for determining the endocardial border on an echocardiogram provided by the present invention can accurately find the endocardial border on an echocardiogram having a mitral valve image, and provides a basis for a computer to automatically calculate the left ventricular ejection fraction of the heart.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention is a method for processing and calculating M-ultrasound cardiogram. Background Art

[0002] Left ventricular ejection fraction (LVEF) is an important indicator for clinical diagnosis of cardiovascular disease and is of great significance for early diagnosis of cardiovascular disease. Echocardiography is the most commonly used, convenient and familiar method for measuring LVEF in clinical practice.

[0003] Echocardiographic methods include M-mode echocardiography (M-echo), which treats the left ventricle as a geometric body of a certain shape, measures the inner diameter of each section, and then calculates the end-diastolic volume and end-systolic volume, and calculates the left ventricular stroke volume and LVEF. At present, in clinical practice, the M-mode ultrasound method is still a manual measurement method to calculate the LVEF index, and the research on automatic detection of LVEF is mostly focused on B-mode echocardiography and nuclear magnetic resonance imaging. However, in the field of handheld ultrasound, manual operation is extremely inconvenient on small-sized terminals such as mobile phones. With the rise of handheld ultrasound, automatic measurement based on M-mode ultrasound is becoming more and more important. For echocardiography, the current mainstream B-mode ultrasound and MRI analysis methods are not suitable for automatic calculation of LVEF by handheld ultrasound, and no literature on automatic measurement of LVEF by M-mode ultrasound has been found. M-mode echocardiography: It uses a single sound beam to scan the heart, and the movement of the heart and large blood vessels is displayed in the form of a curve formed by the change of light points over time. Depending on the placement of the sampling line, the heart base wave group area, mitral valve anterior leaflet wave group area, mitral valve leaflet curve, ventricular wave group, and pulmonary valve wave group can be measured respectively. The position that needs to be measured for automatic calculation of LVEF is the ventricular wave group. The sound waves pass through the right ventricular anterior wall, right ventricular cavity, ventricular septum, left ventricular cavity (and its tendons), and left ventricular posterior wall in sequence. The size of the left and right ventricles and the thickness of the ventricular walls are measured here. The probe is fixed to a certain part of the heart. Due to the regular contraction and relaxation of the heart, the distance between the various layers of the heart tissue and the probe also changes accordingly.

[0004] The computer processes the M-mode echocardiogram by binarizing it, and then recognizes and automatically measures the end-diastolic volume and end-systolic volume displayed on the M-mode echocardiogram, and obtains the ejection fraction according to the formula. Figure 1 As shown: Due to the high density of the left ventricular posterior wall myocardium, the image formed is brighter, such as Figure 1As shown in the white box area. When the image is binarized, the highlighted area causes the binarization threshold to be too large, while the endothelium is a darker area, which ultimately results in inaccurate segmentation of the endothelium area, making it difficult to automatically calculate the end-diastolic volume and end-systolic volume of the heart. In addition, since the mitral valve image is included in the left ventricle of the heart, the mitral valve image is close to the endothelium boundary of the left ventricular septum during imaging, which seriously affects the computer's resolution of the mitral valve image and the endothelium boundary of the left ventricular septum. Therefore, it is necessary to provide a better and more accurate method for obtaining the endothelium boundary of the left ventricular septum.

[0005] At the same time, the left ventricular posterior wall intima boundary is also not clearly separated because other body information is included in the ultrasound image, such as Fig.10 As shown, the two gray boxes are unseparated locations, which makes the calculation more difficult. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a method for determining the endocardial boundary on an echocardiogram.

[0007] The method for determining the endocardial border on an echocardiogram provided by the present invention can accurately find the endocardial border on an echocardiogram with a mitral valve image, thereby providing a basis for a computer to automatically calculate the left ventricular ejection fraction of the heart. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 M-mode echocardiogram pattern;

[0009] Figure 2 M-mode echocardiogram with mitral valve information;

[0010] Figure 3 Schematic diagram of image binarization causing the endocardial border of the left ventricular septum to be lost;

[0011] Figure 4 Black pixel percentage trend chart.

[0012] Figure 5 A schematic diagram of the approximate range;

[0013] Figure 6 Schematic diagram of pixel grayscale values ​​increasing from top to bottom;

[0014] Figure 7 Echocardiograms that eliminate the influence of mitral valve information during binarization;

[0015] Figure 8 Schematic diagram of the left ventricular septal intimal border;

[0016] Fig. 9 The left ventricular septal intimal boundary is a schematic diagram of the connected domain;

[0017] Fig.10 Schematic diagram showing the connection between the left ventricular posterior wall intimal border and the left ventricular internal body;

[0018] Fig.11 Statistical graph of the row position where the white pixel first appears;

[0019] Fig.12 The experimental results obtained in the experiment have information about the mitral valve;

[0020] Fig.13 The experimental results obtained in the experiment without mitral valve information;

[0021] Fig.14 Steps for obtaining the left ventricular septum intimal border and the left ventricular posterior wall intimal border. DETAILED DESCRIPTION

[0022] The present invention is preferably implemented in the form of software code, and the steps of the present invention are edited into software code and installed in a computer with computing functions, for example, in the form of software code, installed in a smart phone or tablet computer, which has the function of recognizing and reading M-mode ultrasound cardiac images.

[0023] Those skilled in the art will appreciate that the exemplary algorithm steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0024] The implementation of the present invention is completed by installing a corresponding computer program in a computer with relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement various required steps. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0025] In the computer of the present invention, the main hardware structure still includes the following main parts: central processing unit, memory, chipset, I / O bus, I / O device, power supply, chassis and related software. The related software refers to the software that can project the M-mode echocardiography (M-echo) image onto the display screen. Of course, the M-mode echocardiography can also be displayed in the form of images, but the computer background can process and calculate the image by itself, and finally only the calculation result is presented on the display.

[0026] The M-mode ultrasound cardiogram of the present invention is in PNG image format; of course, it can also be a DICOM image; a JPEG image, a BMP image, a PNG image, a TIFF image converted from a DICOM image; a JPEG image, a BMP image, a PNG image, a TIFF image.

[0027] The present invention is directed to the processing of M-type echocardiography, which uses a single sound beam to scan the heart and displays the movement of the heart and large blood vessels in the form of a curve formed by the change of light spot groups over time. According to the different placement positions of the sampling line, the heart base wave group area, the mitral valve anterior leaflet wave group area, the mitral valve leaflet curve, the ventricular wave group, and the pulmonary valve wave group can be measured respectively. The position that needs to be measured for automatic calculation of LVEF is the ventricular wave group, and the sound wave passes through the right ventricular anterior wall, the right ventricular cavity, the ventricular septum, the left ventricular cavity (and its tendon), and the left ventricular posterior wall in sequence. The size of the left and right ventricles and the thickness of the ventricular walls are all measured here. The probe is fixedly facing a certain part of the heart. Due to the regular contraction and relaxation of the heart, the distance between the various layers of the heart tissue and the probe also changes accordingly. The formula for measuring LVEF by traditional M-type echocardiography is the Teichholz correction formula, which is an empirical formula proposed after comparing with the LVEF measurement value of biplane angiography. The specific measurement method is: under the guidance of two-dimensional echocardiography, M-type sampling is performed in the standard parasternal left ventricular long-axis section or short-axis mitral valve chordal section, and the sampling line is perpendicular to the ventricular septum and posterior wall as much as possible. The left ventricular end-diastolic diameter (LVDd) and left ventricular end-systolic diameter (LVDs) are measured, and then the left ventricular volume (V) is calculated according to formula (1), where D in the formula is the inner diameter value.

[0028] V=7.0÷(2.4+D)×D 3 (1)

[0029] The inner diameter is in cm, and the left ventricular end-diastolic and end-systolic volumes are in mL. Then LVEF is calculated according to formula (2).

[0030] LVEF(%)=(LVEDV-LVESV)÷LVEDV×100%(2)

[0031] LVEDV and LVESV are the end-diastolic volume and end-systolic volume respectively. Figure 1 shown.

[0032] The technical term grayscale value in the present invention refers to the color depth of a point in a black and white image, and generally ranges from 0 to 255, with white being 255 and black being 0.

[0033] In the embodiment, the pixel points or pixels refer to small squares of an image, each of which has a clear position and an assigned color value, and the color and position of the small squares determine the appearance of the image. A pixel can be regarded as an indivisible unit or element in the entire image.

[0034] In the embodiment, the pixel height refers to the distance between the pixels. A pixel row refers to a pixel row composed of a number of pixels neatly arranged horizontally. A pixel column refers to a pixel column composed of a number of pixels neatly arranged vertically.

[0035] The echocardiogram in the present invention is specifically an M-mode echocardiogram.

[0036] Example 1:

[0037] The end-diastolic volume and end-systolic volume measured on M-mode echocardiography should accurately determine the intimal border of the left ventricular septum and the intimal border of the left ventricular posterior wall on M-mode echocardiography. Fig.14 The left ventricular septum intimal border and the left ventricular posterior wall intimal border can be obtained as follows:

[0038] First, the original image is read in, which is an M-mode echocardiogram, and a virtual center line is set on the echocardiogram, and the virtual center line is roughly located at the same distance from the upper boundary to the lower boundary. The purpose of the virtual center line is to separate the upper and lower images.

[0039] The purpose of separation is that on M-mode echocardiography, the image formed is brighter due to the higher density of the left ventricular posterior wall myocardium, such as Figure 1 As shown in the white box area. When the image is binarized, the highlighted area causes the binarization threshold to be too large, while the endocardium is a darker area, which ultimately results in inaccurate segmentation of the endocardium area. Therefore, the image is divided into two parts, the upper part with the mitral valve image and the left ventricular septum endocardium image, is set with a higher binarization threshold during binarization. The lower part of the left ventricular posterior wall is set with a lower binarization threshold, so different binarization thresholds are set for the upper and lower parts so that they do not interfere with each other.

[0040] Therefore, from the virtual center line, the echocardiogram is divided into two halves: upper and lower halves:

[0041] (1) An upper echocardiogram with a mitral valve image and a left ventricular septum endocardium image; the upper echocardiogram is processed to obtain the left ventricular septum boundary. From an intuitive spatial visual point of view, the upper edge is the upper boundary of the echocardiogram, and the lower edge is the lower boundary of the echocardiogram.

[0042] and: (2) an echocardiographic lower image having the intimal border of the left ventricular posterior wall. The echocardiographic lower image is processed to obtain the intimal border of the left ventricular posterior wall.

[0043] Of course, when the computer processes the upper and lower images, it does not process the upper image first and then the lower image. Instead, it processes them simultaneously, or the order in which the images are processed is not important, nor is there a limit to the order in which the images are processed.

[0044] Example 2:

[0045] The method for setting the virtual center line in Example 1 is:

[0046] 1. First define the upper and lower boundaries of the graph, with the continuous edge close to the mitral valve as the upper boundary of the echocardiogram and the opposite side as the lower boundary.

[0047] 2. Then calculate the "sum" of the grayscale values ​​of all pixels in each pixel row in the M echocardiogram. And find the pixel row with the largest grayscale value sum. Since the density of the left ventricular posterior wall myocardium is higher in the M echocardiogram, the image formed is brighter. Therefore, the pixel row with the largest grayscale value sum must be in the area where the left ventricular posterior wall myocardium density is higher, which is the brighter image formed. Figure 1 The position shown in the box (of course, the row of pixels with the maximum grayscale value, whether there are at least two pixel rows with the same maximum grayscale value, this possibility is basically non-existent on such a fine M-mode echocardiogram. Of course, in order to avoid this possibility, a calculation condition can be added when setting this method: when there are more than two pixel rows with the same maximum grayscale value, take the first pixel row close to the upper boundary that meets the maximum grayscale value condition).

[0048] 3. Starting from the row of pixels with the largest sum of grayscale values, search for the first pixel row with the smallest sum at the upper boundary. Since there is no muscle tissue in the left ventricle of the heart, the scanned image is darker, and the minimum pixel row will appear in the left ventricle of the heart, which is roughly located at the same distance from the upper boundary to the lower boundary in the M echocardiogram.

[0049] The first pixel row with the smallest sum is the virtual center line.

[0050] Example 3:

[0051] Another method of setting the virtual center line in Example 1 is to find the pixel row located at the same distance from the upper boundary to the lower boundary on the echocardiogram, which is the virtual center line. That is, the virtual center line is obtained by directly dividing the center with equal distance from the M echocardiogram.

[0052] Example 4:

[0053] During the process of acquiring cardiac images with M-mode ultrasound, mitral valve information is usually acquired. There is a mitral valve image in the ultrasound cardiac image, such as Figure 2 As shown in the figure, the large box is the area where the mitral valve is located, and the small box is where the mitral valve is connected to the left ventricular septum endothelium, and the connection range is wide. The two become a large connected domain, which will cause the conventional binarization method to be unable to separate the mitral valve from the septum endothelium and obtain the correct septum endothelium boundary. Figure 3 After the original image of the M echocardiogram was binarized, the mitral valve was also a bright area. After the binarization process, it became white, causing the boundary of the left ventricular septum to be lost. The box marked the place where the mitral valve and the ventricular septum were connected. Therefore, the influence of the mitral valve image must be eliminated.

[0054] The method for eliminating the influence of mitral valve image on echocardiography comprises the following steps S01 to S04.

[0055] Step S01:

[0056] On the obtained echocardiogram having at least a mitral valve image and a left ventricular septal endocardium image, the approximate range of the adjacent area of ​​the left ventricular septal endocardium and the mitral valve is determined. The work of determining the approximate range is as follows:

[0057] (S011): On the echocardiogram, the continuous edge close to the mitral valve is the upper boundary of the echocardiogram, and the pixel row below the upper boundary and 30 pixels away from the upper boundary is set as the calculation starting line. From the binarization result, it can be concluded that if the processing starts from the image height (30 pixels away from the upper boundary), the influence of the right ventricle position in the image can be avoided.

[0058] (S012): At the calculation starting point line, downward, count the percentage of black pixels in each pixel row parallel to the calculation starting point line;

[0059] (S013): When the first pixel row with at least 35% black pixels is obtained, it is the upper boundary of the approximate range;

[0060] (S014): Obtaining the lower edge of the approximate range. There are two methods for obtaining the lower edge of the approximate range. Here, both methods are introduced through this embodiment. However, in actual implementation, only one method can be used, and method 1 is preferably used.

[0061] Method 1 for obtaining the lower edge of the approximate range: (I) Continue to count the percentage of black pixels in each pixel row along the upper edge of the range of the mitral lobe pixel block domain downward. When the proportion of black pixels in the pixel row is the largest, the corresponding pixel row is the lower edge of the approximate range.

[0062] (II) When multiple pixel rows satisfying the above condition (I) are obtained, the first pixel row satisfying the above condition (I) that is closest to the upper edge of the range is the lower edge of the approximate range.

[0063] Method 2 for obtaining the lower edge of the approximate range: Continue to count the percentage of black pixels in each pixel row along the upper edge of the range of the mitral valve pixel block domain downwards. When the proportion of black pixels in the row exceeds 80%, and the proportion of black pixels in the subsequent 40mm range is less than 30%, the row in which the proportion of black pixels in the row exceeds 80% is determined to be the lower edge of the approximate range.

[0064] After a lot of experiments, reference Figure 4 At the position corresponding to the middle vertical line, the proportion of black pixels in the row is the largest, and within the subsequent 40 pixels, the proportion of black pixels drops rapidly to less than 30%.

[0065] The area between the upper and lower limits of the range is the approximate range area. Figure 5 , the approximate range marked between the upper limit line of the range and the lower limit line of the range.

[0066] Step S02:

[0067] In the determined approximate range, find the pixel block domain of the adjacent area of ​​the left ventricular septum endothelium and the mitral valve; the pixel block domain is searched row by row from the upper range line to the lower range line within the determined approximate range. The set search conditions are:

[0068] (I) In the same column of pixels, the pixel grayscale values ​​increase from the upper range line to the lower range line for at least ten pixels.

[0069] (II) The first range found to have the above-mentioned feature (I) is the pixel block domain of the adjacent area between the left ventricular septum endothelium and the mitral valve.

[0070] (III) In the step of finding the pixel block domain of the adjacent area between the left ventricular septum endothelium and the mitral valve, when the above search conditions (I) and (II) are met, within the pixel block domain of the adjacent area between the left ventricular septum endothelium and the mitral valve found, only one of the following situations is allowed and is considered as a noise point:

[0071] (1) The grayscale values ​​of no more than two pixels stop increasing compared to the previous pixel; (2) The grayscale value of one pixel decreases compared to the previous pixel.

[0072] refer to Figure 5 , in the approximate range marked between the upper and lower limits, the content in the small box is referenced Figure 6 , the image characteristics of this area are: the pixel gray value increases from top to bottom.

[0073] Step S03:

[0074] The grayscale values ​​of the pixel blocks in the adjacent area are set to 0. This means that these locations belong to the left ventricle, which helps to determine the complete left ventricular septum endothelium.

[0075] Step S04:

[0076] The echocardiogram including the adjacent area whose grayscale value of the pixel block domain is set to 0 is binarized to obtain a binarized echocardiogram in which the adjacent area and the mitral valve area are excluded from the left ventricular septum endocardium. Figure 7 As shown. Figure 3 and Figure 7 , it can be found that the left ventricular septum endothelium has been replenished more completely.

[0077] Example 5:

[0078] The steps to obtain the left ventricular septal intimal border on echocardiography are:

[0079] Step S40:

[0080] Process the echocardiogram, eliminate the influence of the mitral valve image on the echocardiogram, and separate the "mitral valve" and "left ventricular septum endothelium and adjacent areas of the mitral valve" from the left ventricular septum endothelium in the echocardiogram. Refer to Example 4 for the processing method.

[0081] Step S41:

[0082] exist Figure 7 To determine the endocardial border of the left ventricular septum, the region of interest must be delineated first. The region of interest includes the entire endocardial border image of the left ventricular septum; the region of interest is included between the first boundary of the region of interest and the second boundary of the region of interest. The calculation method of the first boundary of the region of interest and the second boundary of the region of interest is as follows:

[0083] (S411) The first boundary of the region of interest:

[0084] On the echocardiogram, the continuous edge close to the mitral valve side is the upper boundary of the echocardiogram, and the pixel row below and 30 pixels away from the upper boundary is set as the first boundary of the region of interest.

[0085] Starting from the point of 30 pixels in height, the division can effectively avoid the influence of the right ventricle position in the image, and the lower boundary of the approximate range of the left ventricular septum endothelium and the mitral valve adjacent can also completely include the left ventricular septum endothelium. Therefore, the first boundary of the region of interest is defined as 30 pixels in height (30 pixels from the upper boundary).

[0086] (S412) Second boundary of the region of interest:

[0087] Two methods are introduced here for determining the second boundary of the region of interest. Both methods are introduced here through this embodiment. However, in actual implementation, only one of the methods can be used, and method one is preferred.

[0088] Method 1: (I) Continue counting the percentage of black pixels in each pixel row from the first boundary of the region of interest downward. When the percentage of black pixels in the pixel row is the largest, the corresponding pixel row is the second boundary of the region of interest.

[0089] (II) When multiple pixel rows satisfying the above condition (I) are obtained, the first pixel row satisfying the above condition (I) that is closest to the first boundary of the region of interest is the second boundary of the region of interest.

[0090] Method 2: Count the percentage of black pixels in each row parallel to the first boundary of the region of interest. When the proportion of black pixels in the row exceeds 80%, and the proportion of black pixels in the subsequent 40mm range is less than 30%, the row in which the proportion of black pixels in the row exceeds 80% is determined to be the second boundary of the region of interest.

[0091] The area between the first boundary of the region of interest and the second boundary of the region of interest is the region of interest.

[0092] Step S43:

[0093] The morphological opening operation was performed on the region of interest to make the binarized echocardiogram smoother; then edge detection was performed to obtain the left ventricular septum endocardial boundary. Figure 8 shown.

[0094] Step S44:

[0095] A closing operation is performed on the obtained left ventricular septum endocardium boundary, and the obtained left ventricular septum endocardium boundary is made into a connected domain. Fig. 9 This is an enlarged view of the boxed part in Figure 8. The clear dividing line shows that the inner membrane boundary has been separated from other connected domains. Figure 8 Find the largest connected domain in the image, which is the inner membrane boundary. By determining the maximum value of the vertical coordinate of the boundary, we can get the required result of the sub-image on the image.

[0096] Example 6:

[0097] refer to Fig.10 , the binarization result of the lower sub-image is as follows Fig.10 As shown in the figure, there are obvious wavy lines. However, the image segmentation result is not good enough. The box part is the position where the left ventricular posterior wall intimal boundary is not separated, so the left ventricular posterior wall intimal boundary and the left ventricular body are connected. When calculating the left ventricular posterior wall intimal boundary, this part must be separated.

[0098] The intimal border of the posterior left ventricular wall is determined on echocardiography as follows:

[0099] Step S11:

[0100] Binarized ultrasound images; the continuous edge close to the mitral valve was calibrated as the upper boundary of the echocardiogram; the continuous edge on the opposite side of the upper boundary was calibrated as the lower boundary of the echocardiogram.

[0101] From the echocardiogram, the distance between the upper boundary and the lower boundary is the virtual midline. Starting from the virtual midline and moving toward the lower boundary, the position of the first white pixel in each column is counted in units of pixel columns. The statistical results are as follows: Fig.11 shown.

[0102] Step S12:

[0103] In the statistically obtained data, if the row where the white pixel appears for the first time is more than 45 pixels away from the upper boundary, the column where it is located is a data error column.

[0104] Step S13:

[0105] Correct the data errors in the column; the method to correct the data errors in the column is:

[0106] (S131): Merge adjacent columns with erroneous data into one erroneous interval;

[0107] (S132): The row where the first white pixel appears in the correct data column adjacent to the left side of the error interval is set as the first reference point, and the row where the first white pixel appears in the correct data column adjacent to the right side of the error interval is set as the second reference point; a straight line connects the first reference point and the second reference point, and the straight line is the left ventricular posterior wall endocardial boundary of the error interval.

[0108] Step S14:

[0109] Morphological opening operation is performed to make the binarized echocardiogram smoother; then edge detection is performed to obtain the intimal boundary of the left ventricular posterior wall;

[0110] Step S15:

[0111] A closing operation is performed on the obtained left ventricular posterior wall intima boundary, and the obtained left ventricular posterior wall intima boundary is made into a connected domain.

[0112] Example 7:

[0113] The implementation method uses PyCharm-Community-2020.1 as the integrated development environment and the OpenCV version is 4.2.0.

[0114] In actual clinical applications, depending on the direction of the probe, M-mode echocardiography may or may not capture the mitral valve information. In order to verify the robustness of this algorithm, experiments were conducted on M-mode ultrasound images with and without the mitral valve. The experimental results are shown in Figure 2. Fig.12 shown. Fig.12 This is the experimental result with mitral valve information, and the calculated LVEF is 55%. Fig.13 This is the experimental result without mitral valve information, and the calculated LVEF is 54%.

[0115] The experimental subjects were 3,000 M-mode echocardiograms, and the average time for automatic measurement of LVEF was 0.6 seconds. Compared with the results calculated manually by clinicians, the accuracy of automatic measurement was 95.6%. This method has a high calculation accuracy and good calculation efficiency.

[0116] The above are various embodiments for implementing the present invention. In the above embodiments, the description of each embodiment has its own emphasis. For the part that is not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0117] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

[0118] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for eliminating the influence of mitral valve image on echocardiography, characterized in that: S01: On an obtained echocardiogram having at least a mitral valve image and a left ventricular septal endocardium image, determine the approximate range of adjacent areas of the left ventricular septal endocardium and the mitral valve; the specific method is: S011: On the echocardiogram, the continuous edge close to the mitral valve side is used as the upper boundary of the echocardiogram, and the calculation starting point line is set in the pixel row below the upper boundary; S012: At the calculation starting point line, downward, count the percentage of black pixels in each pixel row parallel to the calculation starting point line; S013: The first pixel row with at least 35% black pixels is the upper boundary of the approximate range; S014: Obtain the lower boundary of the approximate range in the following manner: (I) Counting the percentage of black pixels in each pixel row along the upper boundary line of the mitral valve pixel block domain downwards, when the percentage of black pixels in the pixel row is the largest, the corresponding pixel row is the lower boundary line of the approximate range; (II) When multiple pixel rows satisfying the above condition (I) are obtained, the first pixel row satisfying the above condition (I) closest to the upper edge of the range is the lower edge of the range of the approximate range; S02: Finding a pixel block region of the adjacent region of the left ventricular septum endothelium and the mitral valve within the approximate range; S03: setting the grayscale value of the pixel block domain in the adjacent area to 0; S04: Binarizing the echocardiogram including the adjacent region whose grayscale value of the pixel block domain is set to 0, to obtain a binary echocardiogram in which the adjacent region and the mitral valve region are excluded from the left ventricular septum endocardium.

2. A method for eliminating the influence of mitral valve image on echocardiography according to claim 1, characterized in that: The calculation starting point line is set at the pixel row below the upper boundary, specifically: the pixel row below the upper boundary and 30 pixels away from the upper boundary is set as the calculation starting point line.

3. A method for eliminating the influence of mitral valve image on echocardiography according to claim 1, characterized in that: The method for finding the pixel block domain of the adjacent area between the left ventricular septum endothelium and the mitral valve is: Within the approximate range, search row by row from the upper range line to the lower range line, and at least meet the following search conditions: (I) In the same column of pixels, the pixel grayscale values ​​increase from the upper range line to the lower range line for at least ten pixels; (II) The first range with the above-mentioned (I) features is the pixel block domain of the left ventricular septum endothelium and the adjacent area of ​​the mitral valve.

4. A method for eliminating the influence of mitral valve image on echocardiography according to claim 3, characterized in that: When the search condition is met, within the pixel block domain of the found left ventricular septum endocardium and the adjacent area of ​​the mitral valve, one of the following situations occurs, which is regarded as a noise point: (I) The grayscale values ​​of no more than two pixels stop increasing compared to the previous pixel; (II) The grayscale value of one pixel decreases compared to the previous pixel.

5. A method for obtaining the intimal border of the left ventricular septum on echocardiography, characterized in that: Follow these steps: S40: Processing the echocardiogram using any one of the methods of claims 1 to 4; S41: delineate the region of interest, which includes the entire left ventricular septum endocardial border image; S43: Perform morphological opening operation on the region of interest to make the binarized echocardiogram smoother; then perform edge detection to obtain the left ventricular septum endocardial boundary; S44: performing a closing operation on the obtained left ventricular septum endocardium boundary, and making the obtained left ventricular septum endocardium boundary a connected domain.

6. The method for obtaining the endocardial border of the left ventricular septum on echocardiography according to claim 5, characterized in that: The calculation method of the region of interest is as follows: S411: On the echocardiogram, the continuous edge close to the mitral valve side is used as the upper boundary of the echocardiogram, and the pixel row below the upper boundary and 30 pixels away from the upper boundary is set as the first boundary of the region of interest; S412: Obtain the second boundary of the region of interest as follows: (I) Count the percentage of black pixels in each pixel row from the first boundary of the region of interest downward. When the percentage of black pixels in the pixel row is the largest, the corresponding pixel row is the second boundary of the region of interest. (II) when multiple pixel rows satisfying the above condition (I) are obtained, the first pixel row satisfying the above condition (I) closest to the first boundary of the region of interest is the second boundary of the region of interest; The area between the first boundary of the region of interest and the second boundary of the region of interest is the region of interest.

7. A method for determining the intimal border of the left ventricular posterior wall on an echocardiogram, characterized in that: S11: Binarized ultrasound image; the continuous edge close to the mitral valve is calibrated as the upper boundary of the echocardiogram; the continuous edge on the opposite side of the upper boundary is the lower boundary of the echocardiogram; On the echocardiogram, the distance between the upper boundary and the lower boundary is the virtual midline. Starting from the virtual midline and moving toward the lower boundary, the row position where the first white pixel appears in each column is counted in units of pixel columns. S12: In the statistically obtained data, if the row where the white pixel appears for the first time is more than 45 pixels away from the upper boundary, the column where the white pixel appears is a data error column; S13: Correct the data errors in the column; S14: performing morphological opening operation to make the binarized echocardiogram smoother; then performing edge detection to obtain the intimal boundary of the left ventricular posterior wall; S15: performing a closing operation on the obtained left ventricular posterior wall intima boundary, and making the obtained left ventricular posterior wall intima boundary a connected domain.

8. A method for determining the intimal border of the left ventricular posterior wall on an echocardiogram according to claim 7, characterized in that: The method for correcting the column data error is: S131: merging adjacent columns with erroneous data into one erroneous interval; S132: Set the row where the first white pixel appears in the correct data column adjacent to the left side of the error interval as the first reference point, and set the row where the first white pixel appears in the correct data column adjacent to the right side of the error interval as the second reference point; connect the first reference point and the second reference point with a straight line, and the straight line is the left ventricular posterior wall endocardial boundary of the error interval.

9. A method for determining the intimal border of the left ventricular septum and the intimal border of the left ventricular posterior wall on an echocardiogram, characterized in that: S91: setting a virtual center line from the echocardiogram, wherein the virtual center line is approximately located at the same distance from the upper boundary to the lower boundary; From the virtual centerline, the echocardiograms were divided into: the upper echocardiogram with the image of the mitral valve and the endocardium of the left ventricular septum; and the lower echocardiogram with the endocardial border of the left ventricular posterior wall; S92: Processing the echocardiogram using any one of the methods of claims 5-6 to obtain the left ventricular septum boundary; The echocardiogram is processed by any one of the methods of claims 7-8 to obtain the intimal border of the left ventricular posterior wall.

10. A method for determining the intimal border of the left ventricular septum and the intimal border of the left ventricular posterior wall on an echocardiogram according to claim 9, characterized in that: To set the virtual centerline, use one of the following methods: (1) Calculate the sum of the grayscale values ​​of all pixels in each pixel row in the echocardiogram; start from the pixel row with the largest sum and look upward for the first pixel row with the smallest sum, which is the virtual center line; (2) From the echocardiogram, find the pixel line located at the same distance from the upper boundary to the lower boundary as the virtual center line.

Citation Information

Patent Citations

  • Method and device for detecting tissue boundaries by use of ultrasonic images

    CN101527047A

  • Cardiac ultrasound imaging method

    CN108013904A

  • Cardiac volume identification analysis system and method

    CN108882917A

  • Ultrasound medical detection devices and imaging method, imaging system and display terminal

    CN109069114A

  • Cardiac medical image processing method, processing device, processing system and medium

    CN110335236A