WO2017206023A1 - Cardiac volume identification analysis system and method - Google Patents

Abstract

Provided are cardiac volume identification analysis method and system, the method comprises: acquiring multiple ultrasound images which are continuously acquired over a predetermined period of time (501, S10); identifying a cardiac section type in the multiple ultrasound image data (501) (S12,S202,S203); identifying a cardiac cycle (S14,S204,S205); identifying the location and shape of a left ventricular endocardium in each of the ultrasound images in one cardiac cycle(S16,S206,S207,S208); calculating a ventricular volume quantification parameter at a time corresponding to each of the ultrasound images according to the location and shape of the left ventricular endocardium, so as to obtain a ventricular volume curve(701,S18,S209); and outputting the ventricular volume curve (701), and/or calculating and outputting clinical parameters characterizing cardiac function on the basis of the ventricular volume curve (701)(S19,S210,S211,S212,S213). This method does not require operations of connecting an ECG signal wire and module in the course of the actual use, which simplifies the user's work and improves the working efficiency.

Description

Heart volume recognition analysis system and method Technical field

The present invention relates to the field of medical technology, and in particular, to a cardiac volume recognition analysis system and method.

Background technique

Ejection Fraction (EF) refers to the percentage of stroke volume in the end-diastolic volume and is one of the important clinical indicators for evaluating left ventricular function. The size of the ejection fraction is related to the contractility of the myocardium. The stronger the myocardial contractility, the greater the stroke volume and the greater the ejection fraction. Under normal circumstances, the left ventricular ejection fraction is >50%. Measuring left ventricular ejection fraction can be accomplished by a variety of means, most of which employ methods based on medical imaging devices. First, the heart image is collected by a medical imaging device such as a CT (Computed Tomography) device, an MRI (Magnetic Resonance Imaging) device, and an ultrasound device. After obtaining an image of a complete cardiac cycle, each is obtained. The left ventricular endocardium is segmented and identified on the frame image, and then the ventricular volume is calculated according to the shape of the endocardium. After obtaining the ventricular volume on each frame of the cardiac cycle, the ventricular volume curve is constructed, and then according to the ventricular volume. The maximum fraction of the curve, the end-diastolic volume (EDV) and the minimum, the end-systolic volume (ESV), is used to calculate the ejection fraction. In the above medical imaging device, echocardiography is a non-invasive and safe diagnostic method that does not require the injection of contrast agents, isotopes or other dyes, and the patient and the doctor are not exposed to radioactive materials. The method is simple and can be repeated many times. Performed at the bedside, each heart chamber can be examined by multi-planar, multi-directional ultrasound imaging to fully evaluate the anatomy and function of the entire heart. The currently used echocardiographic modes for left ventricular ejection fraction measurement include M-Mode mode and B-Mode mode. The left ventricular ejection fraction measurement method based on the M-Mode mode is performed by imaging a line of data on the long axis section of the left ventricle, and then obtaining the EDV and ESV of the ventricular volume by calibrating the maximum inner diameter and the minimum inner diameter of the left ventricle, thereby calculating the shot. Blood score. The left ventricular ejection fraction measurement method based on B-Mode mode obtains the left ventricle two-dimensional image under different sections by imaging the left ventricle, and then recognizes the left ventricular end-systolic and end-diastolic frames according to the image, and then manually corrects the heart. The position of the endometrium was calibrated, EDV and ESV were calculated, and the calculation of the ejection fraction was finally completed. The left ventricular ejection fraction measurement method based on M-Mode has a large dependence on the position of the scan line. It is difficult to collect the standard left ventricular long-axis image for the difference of the heart of different individuals, and it is difficult to obtain the standard measurement left. Scanning line position of indoor path The method of estimating the volume based on the radial line is also not accurate enough. Therefore, the B-Mode mode is a clinically recommended method for measuring ventricular ejection fraction. In B-Mode mode, automatic measurement of ventricular ejection fraction can be achieved at present. The related technology locates the cardiac cycle through ECG signals, identifies different cardiac phases through ECG signals, and then uses image segmentation. Techniques are used to calculate parameters such as ventricular end-systolic volume, end-diastolic volume measurement, and ventricular ejection fraction. Since the current automatic measurement technology of ventricular ejection fraction requires ECG signals, each time the technology is used, it is necessary to connect the ECG signal line and the ECG signal module, which increases the workload of the user and reduces the workload. The user's work efficiency.

Summary of the invention

The invention provides a heart volume recognition analysis system and method, which can reduce the workload of the user and improve the work efficiency.

As an aspect of the present invention, a cardiac volume recognition analysis method is provided, which includes:

Acquiring a plurality of frames of ultrasound images continuously acquired over a predetermined period of time;

Identifying a type of face of the heart in the multi-frame ultrasound image data;

Identify the cardiac cycle;

Identifying the position and shape of the left ventricular endocardium in each ultrasound image of a cardiac cycle based on the above-described slice type;

Calculating a ventricular volume quantitative parameter at a time corresponding to each frame of the ultrasound image according to the position and shape of the left ventricular endocardium, and obtaining a ventricular volume curve;

The ventricular volume curve is output, and/or clinical parameters characterizing cardiac function are calculated and output based on the ventricular volume curve.

Accordingly, as another aspect of the present invention, there is also provided a cardiac volume recognition analysis system, comprising:

An ultrasound image acquisition module, configured to acquire a multi-frame ultrasound image continuously acquired over a predetermined time period;

a facet type identification module for identifying a facet type of the heart in the multi-frame ultrasound image data;

a cardiac cycle recognition module for identifying a cardiac cycle;

a contour acquisition module for identifying the position and shape of the left ventricular endocardium in each frame of the ultrasound image during a cardiac cycle;

A ventricular volume curve generating module for counting the position and shape of the left ventricular endocardium Calculating the ventricular volume quantitative parameter at the time corresponding to each frame of the ultrasound image, and obtaining a ventricular volume curve;

A clinical parameter output module for outputting a ventricular volume curve and/or calculating and outputting clinical parameters indicative of cardiac function from the ventricular volume curve.

In one embodiment of the present invention, there is also provided a cardiac volume recognition analysis system, comprising:

Probe

a transmitting circuit for transmitting an ultrasonic beam to the target object;

a receiving circuit and a beam combining module for obtaining an ultrasonic echo signal;

An image processing module, configured to obtain, according to the ultrasonic echo signal, a multi-frame ultrasound image continuously acquired over a predetermined time period, identify a type of the heart surface in the ultrasound image, identify a cardiac cycle, and identify each frame of the ultrasound image in a cardiac cycle The position and shape of the left ventricle endocardium, and calculate the ventricular volume quantitative parameter at the time corresponding to each frame of the ultrasound image to obtain a ventricular volume curve;

A display for displaying the above ultrasound image and ventricular volume curve, marking the position and shape of the left ventricular endocardium, and displaying the type of the above-mentioned section.

The present invention provides a heart volume recognition analysis system and method. The invention utilizes the characteristics of cardiac motion and image processing technology to replace the prior art recognition of the automatic cardiac cycle based on the electrocardiographic signal and the determination of the cardiac motion phase. In the actual use process, there is no need to connect the ECG signal wires and modules, which simplifies the user's workload and improves work efficiency.

DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below. Obviously, the drawings in the following description are only It is a certain embodiment of the present invention, and other drawings can be obtained from those skilled in the art without any creative work.

1 is a schematic structural view of an ultrasonic imaging apparatus provided by the present invention;

2 is a schematic flow chart of an embodiment of a method for analyzing cardiac volume according to the present invention;

Figure 3 is a more detailed flow chart of step S12 of Figure 2;

4 is a schematic diagram of identifying a cardiac cycle in an embodiment of step S14 of FIG. 2;

FIG. 5 is a schematic diagram of obtaining an endocardial contour based on edge extraction of feature points in an embodiment of step S16 of FIG. 2; FIG.

6 is a schematic diagram of a machine learning network obtaining an endocardial contour in another embodiment of step S16 in the figure;

Figure 7 is a schematic diagram of a ventricular volume curve obtained in an embodiment of step S18 of Figure 2;

Figure 8 is a schematic flow chart of one embodiment of the present invention;

9 is a schematic structural view of an embodiment of a cardiac volume recognition and analysis system according to the present invention;

10 is a schematic structural view of an embodiment of a face type identification module of FIG. 9;

11 is a schematic structural view of another embodiment of the face type identification module of FIG. 9;

12 is a schematic structural diagram of an embodiment of the central dynamic period identification module of FIG. 9;

13 is a schematic structural diagram of another embodiment of the central dynamic period identification module of FIG. 9;

FIG. 14 is a schematic structural view of an embodiment of the contour obtaining module of FIG. 9.

detailed description

Due to the noise, artifacts, and structural complexity of some anatomical tissues, there are different types of facets for the corresponding facet images. For example, for heart ultrasound images, there are usually different types of facets, such as: There are various types of face types, such as apical two-chamber heart, apical four-chamber heart, and the position and shape of the endocardium of the left ventricle are different in different types of heart-face types. Therefore, for cardiac ultrasound image data with multiple types of facets, if the fixed shape model is used for image segmentation operation when automatically identifying the endocardial region, image extraction errors must exist, and it is difficult to distinguish multiple heart slices in automatic recognition. The difference in position and shape of the intimal endocardium makes the segmentation and recognition of the endocardium inaccurate, resulting in errors in the measurement results.

At present, automatic measurement of ventricular ejection fraction has been achieved. The related technology locates the cardiac cycle through the ECG signal, recognizes different cardiac phases through ECG signals, and then uses image segmentation technology to achieve ventricular end-systolic volume. Measurement of end-diastolic volume and calculation of ventricular ejection fraction. The measurement of ventricular end-systolic volume and end-diastolic volume can be performed by segmenting and identifying the endocardium on the ultrasound images of ventricular end-systolic and end-diastolic, and then calculating the ventricular volume using a ventricular volume calculation formula such as the Simpson method. It can be seen that for the B-Mode mode, the accuracy of the segmentation and recognition of the endocardium directly determines the accuracy of the ventricular volume measurement. Therefore, achieving accurate segmentation and recognition of the endocardium is critical to improving the accuracy of system measurements.

FIG. 1 provides a schematic structural diagram of a system of an ultrasound image acquisition device. This article to get the heart The structure of the system is described in detail by ultrasound. As shown in FIG. 1, the apparatus for performing ultrasound imaging on a target area according to an embodiment of the present invention includes: a probe 1, a transmitting circuit 2, a transmitting/receiving selection switch 3, a receiving circuit 4, a beam combining module 5, a signal processing module 6, and an image. Processing module 7 and display 8. The transmitting circuit 2 transmits the delayed-focused ultrasonic pulse having a certain amplitude and polarity to the probe 1 through the transmitting/receiving selection switch 3. The probe 1 is excited by the ultrasonic pulse to transmit ultrasonic waves to a target area (not shown in the figure, such as cardiac tissue) of the body to be tested, and receive ultrasonic echoes with tissue information reflected from the target area after a certain delay. And re-convert this ultrasonic echo into an electrical signal. The receiving circuit receives the electrical signals generated by the conversion of the probe 1 to obtain ultrasonic echo signals, and sends the ultrasonic echo signals to the beam combining module 5. The beamforming module 5 performs processing such as focus delay, weighting, and channel summation on the ultrasonic echo signals, and then sends the ultrasonic echo signals to the signal processing module 6 for related signal processing. The ultrasonic echo signals processed by the signal processing module 6 are sent to the image processing module 7. The image processing module 7 performs different processing on the signals according to different imaging modes required by the user, obtains image data of different modes, and then forms ultrasonic images of different modes by logarithmic compression, dynamic range adjustment, digital scan conversion, etc., such as B image, C image, D image, and the like. The ultrasound image generated by the image processing module 7 is sent to the display 8 for display. For example, an ultrasound image of systolic end-systolic and end-diastolic phases can be simultaneously displayed on the display interface, and the endocardial contour can be outlined on the ultrasound image. The ultrasound images of the systolic end-systolic and end-diastolic phases for simultaneous display may be standard ultrasound section images (eg, apical two-chamber heart, apical four-chamber heart, etc.), or ultrasound section images corresponding to any section selected by the user.

In addition, the system shown in FIG. 1 further includes an operation control module 9 through which the device user can input a control command on the display interface, for example, inputting a modified contour mark, annotating mark text, and performing mode switching on the ultrasonic image. Wait for operational instructions.

Based on the above system configuration, a method and system for heart volume recognition analysis is provided in an embodiment of the present invention. Can reduce the user's workload and improve work efficiency.

2 is a schematic flow chart of an embodiment of a cardiac volume identification analysis method according to the present invention; as shown in the figure, the method includes:

Step S10: Acquire a multi-frame ultrasound image continuously acquired over a predetermined time period, and the multi-frame ultrasound image may include a continuously acquired multi-frame ultrasound heart image, and may also include progressively from a B-mode ultrasound heart movie, an M-mode super heart. One of the movies, etc., or a combination of both. For example, in one implementation of the invention, it is an ultrasound heart movie of at least 3 seconds in length with an image frame rate of no less than 25 frames per second, so the predetermined time may be greater than 3 seconds. The multi-frame ultrasound image in this embodiment may It is ultrasonic image data acquired in real time, and may also be ultrasonic image data obtained by buffering or remote transmission. If it is the ultrasound image data acquired in real time, before step S10, the method further includes:

First, an ultrasonic beam is emitted to a heart region of a target object; then, an ultrasonic echo signal is obtained, and a multi-frame ultrasonic image continuously acquired over a predetermined period of time is obtained based on the ultrasonic echo signal.

In step S12, the type of the face of the heart in the multi-frame ultrasound image data is identified. In one embodiment of the invention, the type of section includes a standard section of the target object in medical anatomy or ultrasound imaging, for example, a type of section for cardiac tissue including, but not limited to, a four-chamber heart, a two-chamber heart, and the like. Of course, the above-mentioned aspect type is not limited to the standard cut surface, and may also include a custom cut surface type. For example, the custom cut surface type may be an ultrasonic cut surface image obtained by the user selecting an arbitrary direction to cut the target object. The ultrasound image herein can be obtained, but not limited to, using only the system shown in Figure 1 above. In this step, the identified type of the facet can be displayed on the display interface.

In step S14, the cardiac cycle is identified. In one of the embodiments of the present invention, the multi-frame ultrasound image data is analyzed to obtain a cardiac cycle. After the determination of the type of the heart slice is completed, the multi-frame image in the ultrasound cardiac film needs to be analyzed to identify the cardiac cycle; the specific method of identifying the cardiac cycle will be described in detail below. An ultrasound movie or an ultrasound movie file herein can be understood as a representation or storage form of a multi-frame ultrasound image that is continuously acquired over a predetermined period of time.

Step S16, identifying the position and shape of the left ventricular endocardium in each frame of the ultrasound image within a cardiac cycle. The location includes the identified location coordinates, orientation information, etc. displayed by the endocardium in the ultrasound image, which may include coordinate position information for one or more pixel points. The shape includes model parameters for simulating the overall shape of the endocardium, the model parameters including basic parameters and deformation parameters for expressing simulated structures such as circles, ellipses, etc., and deformation parameters including distortion parameters, scaling parameters, manual or automatic adjustment parameters , stretching parameters, and more. Of course, the shape here can be represented by model parameters, and the identified position can also be used to obtain coordinate position information of a set of discrete or continuous pixels for characterizing the shape of the identified left ventricular endocardium.

Step S18, calculating a ventricular volume quantitative parameter at a time corresponding to each frame of the ultrasound image according to the position and shape of the left ventricular endocardium, and obtaining a ventricular volume curve.

Step S19, calculating a clinical parameter characterizing cardiac function according to the ventricular volume curve described above, and outputting; and/or outputting a ventricular volume curve. Among them, finding the maximum value on the ventricular volume curve is the end-diastolic volume (EDV) in the current cardiac cycle, and finding the minimum value is the end-systolic volume (ESV) of the current cardiac cycle. According to EDV and ESV, the ejection fraction (EF) of the left ventricle can be calculated. Important clinical parameters that characterize cardiac function, such as stroke volume and cardiac output. These clinical parameters can be output to the display for display, which can be displayed in a text display manner. It can also be output by voice prompts.

Each step in Figure 2 will be described in more detail below.

FIG. 3 is a more detailed flowchart of step S12 of FIG. 2; in this embodiment, the step S12 includes:

Step S120, identifying a position of the interventricular space in the ultrasound image;

Step S122, rotating the ultrasound image according to the position of the chamber interval, so that the long axis direction of the left ventricle in the ultrasound image is vertical;

Step S124, translating the ultrasound image to adjust the position of the left ventricle in the ultrasound image to the center of the image. The above steps S120 to S124 can be regarded as a normalization process of the ultrasonic image. In the normalized process, the ultrasound image in step 120 may include a frame image, and may also include each frame in the partial multi-frame image, and may also include each frame of the above-described multi-frame image.

In step S126, one or more frames of ultrasound image data are mapped to the feature space, and the feature space is constructed by extracting features in the training set image.

It can be understood that, in some examples, the training set image includes at least a cardiac ultrasound image corresponding to various types of cut surfaces, such as a two-chamber view, a four-chamber view, and the like. The feature space can be constructed by extracting the features in the training set image. The feature extraction can be performed by principal component analysis, or the HAAR feature of the image can be extracted, or the anatomical structure features of the heart can be extracted to construct the feature space. .

Step S128, comparing the projection of the ultrasonic image in the feature space with the projection of the training image of the known slice type in the feature space, and determining the type of the slice of the one or more frames of the ultrasound image. Specifically, the projection of the image to be classified in the feature space after the normalization (ie, the recognition, rotation, and translation in the above steps S120 to S124) is compared with the projection of the training image in the feature space, and the nearest neighbor may be adopted. Or the K-nearest neighbor method classifies and identifies the image types to be classified by the classified image.

It can be understood that, in one embodiment of the present invention, the slice type of one frame of the ultrasound image continuously acquired in the predetermined time period can be equated with the above-mentioned continuous acquisition in the predetermined time period. The type of the slice corresponding to the frame ultrasound image; or in other examples, based on the type of the slice of the ultrasound image of each frame in the multi-frame ultrasound image described above, The type of the slice corresponding to the frame ultrasound image, for example, the multi-frame image can be identified, and then the voting facet type of the movie is finally determined by the voting algorithm.

Therefore, in one embodiment of the present invention, a feature of one or more frames of the ultrasound image in the multi-frame ultrasound image is compared with a feature of a training image of a known slice type to obtain the multi-frame ultrasound image. The type of aspect of the data. The features in this embodiment may include a positional relationship of a segmentation region (such as an anatomical structure) in an image, an image pixel value, an image pixel value distribution, a shape and a size of a distribution region (such as an anatomical structure) in the graphic, and the like. Information that can be used in the image to extract image feature recognition can be included in the features in this embodiment.

In some examples, in the above step S14, the step of identifying the cardiac cycle can be specifically implemented by the following method:

First, extracting feature values of each frame of the ultrasound image according to the multi-frame ultrasound image continuously acquired in the predetermined time period to generate a characteristic curve;

Then, the above characteristic curve is periodically analyzed to identify the cardiac cycle of the target.

In one embodiment of the present invention, the feature value of each frame image is a similarity coefficient, and the feature curve is a similarity coefficient curve.

Specifically, the characteristic curve may be an image similarity curve. The similarity curve is generated by selecting a certain frame in the loaded cardiac ultrasound film as a standard frame (501), and calculating a similarity coefficient between each frame image and the standard frame in the loaded cardiac ultrasound movie, and generating Similarity coefficient curve (503). The method for calculating the similarity coefficient of each frame image and the standard frame in the loaded cardiac ultrasound movie file may be: first, calculating each pixel point and standard frame on each frame in the cardiac ultrasound movie file. Corresponding to the sum of the absolute values of the difference values of the gray values of the pixel points, the summation value is used as a similarity coefficient for describing the degree of similarity between the two frames of images; second, the image of each frame in the cardiac ultrasound movie file is seen As a matrix, the value of each pixel in the image is taken as the element value of the matrix, and the positive correlation coefficient between the matrices is calculated as a similarity coefficient describing the degree of similarity between the two frames of images. In addition, since the number of pixels of the original image is large, it takes more time to calculate the similarity coefficient, and the original image can be downsampled, the original image is reduced to an appropriate scale, and the calculation is reduced without losing the image information. The time required for the similarity factor. In addition, the locality region in the image may be selected to calculate the similarity coefficient, and the local region in the selected image may be: a ventricular septal region in the cardiac ultrasound image, a mitral valve region in the cardiac ultrasound image, and the like. Selecting local regions in the image further reduces the time required to calculate the similarity coefficients between images.

In one embodiment of the present invention, in the extracting the feature values of each frame of the ultrasound image, the above The feature values of each frame of image include image measurement values such as tissue anatomical measurements, which are image measurement values of the image measurement values as a function of time.

Specifically, the characteristic curve may also include an image measurement curve of the tissue anatomy measurement value as a function of time, including but not limited to: left ventricular long axis length, left ventricular area, left ventricular volume, or right ventricular volume, etc. Structural measurements.

The image measurement curve is a plot of tissue anatomical measurements over time, such as a curve of left ventricular volume over time, a curve of left ventricular area over time, a curve of left ventricular long axis length over time, or right ventricular volume. Curves that change over time, and so on.

For example, in the process of generating an image measurement parameter curve, a preliminary contour of the left ventricle is first generated. The preliminary contour can roughly describe the morphological changes of the left ventricle, but may not require pixel level accuracy. The method of generating the preliminary contour can be obtained by locating the left ventricle according to the feature points such as the apex and the mitral annulus, or by extracting the boundary of the endocardium through a low-resolution image, thereby obtaining the above-mentioned measured values, such as the left ventricular long axis. Length, left ventricular area, left ventricular volume, etc. Based on the obtained measured value as a function of time, an image measured value curve is obtained as a characteristic curve for obtaining a cardiac cycle.

In the above step S14, after the characteristic curve is obtained, the cardiac cycle can be identified based on the characteristic curve. Specifically, the identification of the cardiac cycle can be performed in several ways as follows:

First: Select one frame of ultrasound image as the standard frame, and identify the complete period of the standard frame according to the characteristic curve. The method of selecting a standard frame may be randomly selecting a frame image as a standard frame in the loaded ultrasound movie, or randomly selecting one frame image as a standard frame in the loaded multi-frame ultrasound image, or selecting a heart contraction or a certain frame image in the diastolic process as a standard frame;

The local extremum is searched in the local area near the time when the standard frame is located, and another frame ultrasound image corresponding to the local extremum is determined; the time corresponding to the standard frame is taken as the start and end points of the time, and a cardiac cycle is obtained. Specifically, determining a time period in the ultrasound movie or the multi-frame ultrasound image corresponding to the standard frame and the determined another frame image as a start point and an end point of the time, the heart movie file or the partial multi-frame in the time period The ultrasound image can be determined as a cardiac cycle. As shown in Fig. 4, such an example is shown. Wherein, the local extremum is searched in a local area near the time when the standard frame is located, and the ultrasound image 402 corresponding to the local extremum is determined. The time in the ultrasonic movie corresponding to the standard frame 401 and the determined another frame image 402 is used as the start and end points of time to determine a time period 404, and the cardiac film file in the time period can be determined as a cardiac cycle.

Second: the method of identifying the cardiac cycle can also be based on the period of the characteristic curve, and then according to the standard The position of the quasi-frame and the length of the period determine a complete cardiac cycle. The method of calculating the period of the characteristic curve may be to transform the similarity curve in the time domain to the frequency domain by using the Fourier transform, obtain the spectrum of the characteristic curve, find the peak in the spectrogram, and then calculate the period of the characteristic curve according to the peak value of the spectrum. The method of calculating the period of the characteristic curve may also be to calculate an autocorrelation coefficient curve of the characteristic curve, and determine the period according to the position of the peak of the autocorrelation coefficient curve.

Wherein, in step S16, the step of identifying the position and shape of the left ventricular endocardium in each frame of the ultrasound image in a cardiac cycle may include:

Selecting a frame image as a key frame, and performing image segmentation on the left ventricular endocardium of the key frame based on the above-mentioned slice type to obtain a standard endocardial segmentation model;

The position and shape of the left ventricular endocardium on each ultrasound image within a cardiac cycle is identified according to the standard endocardial segmentation model described above.

In one embodiment of the present invention, image segmentation of the left ventricular endocardium of the above key frame based on the above-mentioned aspect type can be performed by using a conventional method for edge extraction based on feature points, that is, edge extraction based on feature points, and key frames are generated. The position and shape of the endocardium; as shown in FIG. 5, a schematic diagram showing the endocardial contour obtained by edge extraction based on the feature points in one embodiment of step S16; in the case of a heart containing a two-chamber heart or a four-chamber heart-cut surface, etc. In the standard section of the ultrasound heart image 501, the initial shape and position 502 of the endocardium is generated based on the detected positions of the two key points 503, 504 of the mitral annulus, and then several are detected near the initial shape of the endocardium. The maximum point 505 of the partial image gradient is then the position and shape of the endocardium based on the detected maximum point of the local image gradient and the two key points of the mitral annulus.

It will be appreciated that in one embodiment of the present invention, endocardial segmentation and recognition may also employ a machine learning approach that identifies the overall model and identifies the location and shape of the key frame endocardium. Implementations may be, but are not limited to, deep learning based left ventricular segmentation algorithms, such as convolutional neural network (CNN) plus linear regression, as shown in FIG. Input a frame image, CNN uses the convolution kernel to extract the feature layer-by-layer convolution, and finally uses linear regression to estimate the final contour on the extracted features. The model parameters of CNN and linear regression have been obtained by training set training.

Of course, in other embodiments of the invention, endocardial segmentation and recognition may also be a combination of both conventional methods and machine learning. The traditional method can be used to extract the key points and locate the left ventricle. On this basis, the endocardial extraction can be accurately performed by machine learning. It is also possible to identify the endocardium based on the overall model and combine the traditional boundary extraction algorithm to further optimize the boundary.

After obtaining the endocardium of the key frame, according to the above-mentioned standard endocardial segmentation model, the steps of identifying the position and shape of the left ventricular endocardium on each frame of the ultrasound image in a cardiac cycle may be performed by the following methods. :

First: After obtaining the endocardium of the key frame, the same segmentation algorithm can be used to segment the ultrasound images other than the key frame in the cardiac cycle by frame-by-frame to obtain the left ventricle of each other frame in the cardiac cycle. The position and shape of the membrane.

Second: based on the position and shape of the segmented key frame endocardium, the endocardial motion tracking process is performed on other ultrasound images except the key frame in the cardiac cycle to obtain the left ventricular heart on each frame of the ultrasound image. The location and shape of the intima.

Specifically, the method of motion tracking may be based on block matching. For example, the segmented endocardial curve is discretized into a plurality of tracking points, and an image in a neighborhood of a certain size centered on the tracking point on the current image is used as an initial block, and the tracking point is centered on the next frame image. The search area is constructed in the neighborhood of the size, and the target block matching the initial block is selected according to the gray similarity in the search area, and then the center of the target block is used as the position of the tracking point on the next frame image, and each track of the current frame is obtained. The position and shape of the endocardium on the image of the next frame can be obtained after the position on the next frame image. Performing the above operations for each frame of the determined one cardiac cycle can obtain the position and shape of the endocardium on each frame of the cardiac cycle. Motion tracking can also be based on optical flow methods or other tracking algorithms.

It can be understood that, in one embodiment of the present invention, the method of generating the endocardium of other frame images in the cardiac cycle after obtaining the endocardium of the key frame may also be a combination of segmentation and motion tracking. The endocardium of each frame is combined with the segmentation and tracking double results for intelligent fusion to obtain the optimal boundary. The method of fusion may be a linear combination of the results of the two, such as direct averaging of the position; or the intelligent selection of the local area according to the confidence of the results of the two parties, and then the fusion of the results after the selection.

In step S18, according to the position and shape of the left ventricular endocardium, the ventricular volume quantitative parameter at the time corresponding to each frame of the ultrasound image is calculated, and a ventricular volume curve is obtained;

Specifically, the position and shape of the endocardium on each frame of the image is obtained, and the quantitative analysis of the left ventricle can be performed accordingly. Among them, quantitative analysis parameters include, but are not limited to, left ventricular diameter, area, volume, and the like. The inner diameter may include the length of the inner diameter in the long axis direction, which is generally defined as the distance from the apex to the midpoint of the mitral annulus, and may also include the length in the direction of the minor axis, such as the end of the mitral annular chamber from one end to the end of the left ventricular side wall. the distance. The left ventricular area can be calculated by the method of accumulating pixels inside the heart chamber. Left ventricular volume It can be calculated by the Simpson method or the area length method. The volume parameter can be further divided into the left ventricular total volume (referred to as the left ventricular volume) and the left ventricular local volume (local volume). The local volume is the volume value in each segment of the myocardium. The segmentation can be performed in 16 segments or 17 sections defined by the American Society of Echocardiography (ASE) and the American Heart Association (AHA). Segment model. According to the above scheme, the ventricular quantitative parameter of the time corresponding to each frame image in the current cardiac cycle is calculated, and the ventricular quantitative analysis parameter variation curve in the current cardiac cycle can be obtained. As shown in Fig. 7, a schematic diagram of a left ventricular volume curve is shown.

At step S19, clinical parameters characterizing cardiac function are calculated based on the ventricular volume curve described above, and output. It can be understood that, in step S19, the ultrasound image corresponding to the time at which the maximum value and the minimum value on the ventricular volume curve are located can be simultaneously displayed, and the position and shape of the left ventricular endocardium are marked. Specifically, the time corresponding to the EDV in the ventricular volume curve is used as the end diastolic time of the current cardiac cycle, and a frame image of the end-diastolic phase in the cardiac ultrasound movie is located, and then the time corresponding to the ESV in the ventricular volume curve is used as the current cardiac motion. At the end of the systolic phase of the cycle, an image of the end of systole in the cardiac ultrasound film is located, and then the end-diastolic and end-systolic images of the current cardiac cycle and the endocardial position and shape are displayed on the display interface.

Referring to FIG. 7, finding the maximum value 702 in the ventricular volume change curve 701 can obtain the corresponding frame of the current cardiac cycle end-diastolic volume and end-diastolic period in the cardiac cycle, and finding the minimum value 703 can obtain the current cardiac cycle end-systolic volume and end-systolic period. Corresponding frame within the cardiac cycle. According to the ventricular volume at the end of diastole and end-systolic, it is possible to calculate the ejection fraction of the left ventricle, stroke volume and cardiac output, and other important clinical indicators that characterize cardiac function.

It can be understood that, in an embodiment of the present invention, in step S12, the method further includes: prompting the user to modify the displayed aspect type; and when the user inputs the modified aspect type, updating the aspect type.

In one embodiment of the invention, the method further comprises: displaying the calculated calculated ventricular volume curve, and/or clinical parameters indicative of cardiac function.

In one embodiment of the present invention, after the step S14, the method further includes: switching to the manual input mode after the failure of the cardiac cycle recognition, for acquiring a cardiac cycle manually input by the user.

In an embodiment of the present invention, the foregoing step S16 includes:

Display the above key frames;

Prompt the user to manually perform the image segmentation process of the left ventricular endocardium in the above key frames. Pre

The above-described standard endocardial segmentation model is obtained based on the adjustment or input result of the user on the above key frames.

In an embodiment of the present invention, after the step S16, the method further includes:

The segmentation result of the position and shape of the left ventricular endocardium is determined. When it is determined that the segmentation result is incorrect, the jump to the manual input mode is used to prompt the user to input the position and shape of the left ventricular endocardium on the ultrasound image.

In an embodiment of the present invention, the method further includes:

Prompt the user to confirm and/or correct the current display and output results.

Based on the user's confirmation and/or correction, an output report is formed, which includes at least: a time corresponding to the end of diastole, a time corresponding to the end of systole, a position and shape of the left ventricular endocardium marked on the ultrasound image, and a function of characterizing the heart. One of the clinical parameters.

Accordingly, one of the embodiments of the present invention is provided in FIG. In this embodiment, a more detailed flow chart is shown. The ultrasound image is first loaded (step S201). After loading the ultrasound image, the system identifies and determines the type of the heart slice in the ultrasound image (step S202), and displays the result on the display interface. If the current slice type is determined to be incorrect, the determination result is modified by the user ( Step S203). After the determination of the type of the heart slice is completed, the system analyzes the multi-frame image in the ultrasound cardiac film to realize the recognition of the cardiac cycle (step S204). If the system automatically recognizes the failure, the cardiac cycle is manually specified by the user (step S205). After the recognition of the cardiac cycle is completed, the system automatically selects one frame of image, performs automatic segmentation of the endocardium (step S206), and determines the automatic segmentation result. If the system determines that there is an error in the automatically recognized endocardium, the user manually inputs the endocardium on a certain frame image (step S207). After automatically identifying the accurate endocardium or manually entering the endocardium by the user, the system uses the position and cardiac cycle information of the endocardium to identify the position of the left ventricular endocardium on each frame of the cardiac cycle and Shape (step S208). Based on the position and shape of the left ventricular endocardium on each frame of image, the system calculates the ventricular volume and other quantitative analysis parameters at the time corresponding to each frame of image. The ventricular volume curve is obtained based on the ventricular volume at several moments in one cardiac cycle (step S209). Finding the maximum value on the ventricular volume curve is the end-diastolic volume (EDV) in the current cardiac cycle, and finding the minimum value is the end-systolic volume (ESV) of the current cardiac cycle. According to EDV and ESV, it is possible to calculate the left ventricular ejection fraction (EF), stroke volume and cardiac output, and other important clinical parameters that characterize cardiac function. At the same time, the time corresponding to the EDV in the ventricular volume curve is taken as the current cardiac motion. At the end of diastole of the cycle, locate a frame of end-diastolic images in the echocardiogram of the heart, and then use the time corresponding to the ESV in the ventricular volume curve as the end-systolic time of the current cardiac cycle, and locate an image of the end-systolic phase in the echocardiogram of the heart. Then, the end-diastolic and end-systolic images of the current cardiac cycle and the endocardial position and shape are displayed on the display interface (step S210). The user needs to judge the current result (step S211), and if the user approves the current result, the user can input the current result into the final report (step S213). If the user does not recognize the current result, the user needs to manually modify the current result (step S212), the modified content includes: the end of diastole or end-systolic period and the location and shape of the endocardium at the end of systole or end-diastolic. After the user modifies the result, the user can input the current result into the final report (step S213).

FIG. 9 is a schematic structural diagram of an embodiment of a cardiac volume recognition and analysis system according to the present invention; in this embodiment, the cardiac volume recognition and analysis system 1 includes:

The ultrasound image acquisition module 10 is configured to acquire a multi-frame ultrasound image continuously acquired over a predetermined time period;

The facet type identification module 11 identifies the type of the face of the heart in the multi-frame ultrasound image data;

a cardiac cycle identification module 12 for identifying a cardiac cycle;

a contour obtaining module 13 configured to identify a position and a shape of a left ventricular endocardium in each frame of ultrasound images in a cardiac cycle based on the above-described slice type;

The ventricular volume curve generating module 14 is configured to calculate a ventricular volume quantitative parameter at a time corresponding to each frame of the ultrasound image according to the position and shape of the left ventricular endocardium, and obtain a ventricular volume curve;

a clinical parameter output module 15 for outputting a ventricular volume curve, and/or calculating and outputting clinical parameters characterizing cardiac function according to the ventricular volume curve;

The display marking module 16 is configured to display an ultrasound image corresponding to the time at which the maximum and minimum values on the ventricular volume curve are located, and to mark the position and shape of the left ventricular endocardium.

FIG. 10 is a schematic structural diagram of an embodiment of a facet type identification module in FIG. 9;

The above-described aspect type identifying module 11 further includes:

The facet type display module 110 is configured to display the identified type of the facet.

The facet type modification prompting module 111 is configured to prompt the user to modify the displayed type of the facet;

The facet type update module 112 is configured to update the facet type when the user inputs the corrected facet type.

As shown in FIG. 11, the structure of another embodiment of the facet type identification module in FIG. In the embodiment, the aspect type identification module 11 further includes:

a position recognition module 113, configured to identify a position of the interventricular space in the ultrasound image;

a rotation processing module 114, configured to rotate the ultrasound image according to the position of the chamber interval, so that the long axis direction of the left ventricle in the ultrasound image is vertical;

The translation processing module 115 is configured to translate the ultrasound image to adjust the position of the left ventricle in the ultrasound image to the center of the image.

The feature space mapping module 116 is configured to map one or more frames of ultrasound image data to the feature space, where the feature space is constructed by extracting features in the training set image;

The comparison determining module 117 is configured to compare the projection of the ultrasound image in the feature space with the projection of the training image of the known slice type in the feature space, and determine the type of the slice of the one or more frames of the ultrasound image.

As shown in FIG. 12, it is a schematic structural diagram of an embodiment of the central dynamic period identification module of FIG. 9. In this embodiment, the cardiac cycle identification module 12 includes:

The characteristic curve generating module 120 is configured to extract feature values of each frame image according to the multi-frame ultrasound image continuously acquired in the predetermined time period to generate a characteristic curve;

The first identification module 121 is configured to perform periodic analysis on the feature curve to identify a cardiac cycle of the target.

FIG. 13 is a schematic structural diagram of another embodiment of the central dynamic period identification module of FIG. 9;

In this embodiment, the cardiac cycle identification module 12 includes:

The standard frame selection module 122 is configured to select a frame of the ultrasound image as a standard frame;

The second identification module 123 is configured to search for a local extremum in a local area near the time when the standard frame is located, and determine an ultrasound image corresponding to the local extremum; a time corresponding to the standard frame as a start point and an end point of the time, obtain a Cardiac cycle.

FIG. 14 is a schematic structural view of an embodiment of the contour obtaining module of FIG. 9. In this embodiment, the contour obtaining module 13 includes:

The key frame contour obtaining module 130 is configured to select a frame image as a key frame, and perform image segmentation on the left ventricular endocardium of the key frame based on the slice type to obtain a standard endocardial segmentation model;

The other frame contour obtaining module 132 is configured to identify the position and shape of the left ventricular endocardium on each frame of the ultrasound image within a cardiac cycle in accordance with the standard endocardial segmentation model described above.

The key frame contour obtaining module 130 obtains the position and shape of the key frame endocardium by any of the following methods:

Edge extraction based on feature points to generate the position and shape of the key frame endocardium; and/or

Using the machine learning method, the position and shape of the key frame endocardium are identified based on the overall model.

Wherein, the other frame contour obtaining module 132 obtains the position and shape of the left ventricular endocardium on each frame of the ultrasound image in any of the following ways:

According to the above-mentioned standard endocardial segmentation model, the ultrasound images other than the key frame in the cardiac cycle are segmented frame by frame to obtain the position and shape of the left ventricular endocardium on each frame of the ultrasound image; or

Based on the position and shape of the segmented key frame endocardium, the endocardial motion is tracked on other ultrasound images except the key frame in the cardiac cycle to obtain the left ventricular endocardium on each frame of the ultrasound image. Location and shape.

For further details, reference may be made to the foregoing description of FIGS. 1 through 8, which are not described in detail.

Based on the system structure shown in FIG. 1, in one embodiment of the present invention, a heart volume recognition and analysis system is further provided, which includes:

Probe

a transmitting circuit for transmitting an ultrasonic beam to the target object;

a receiving circuit and a beam combining module for obtaining an ultrasonic echo signal;

An image processing module, configured to obtain, according to the ultrasonic echo signal, a multi-frame ultrasound image continuously acquired over a predetermined time period, identify a type of the heart surface in the ultrasound image, identify a cardiac cycle, and identify each frame of the ultrasound image in a cardiac cycle The position and shape of the left ventricle endocardium, and calculate the ventricular volume quantitative parameter at the time corresponding to each frame of the ultrasound image to obtain a ventricular volume curve;

A display for displaying the above ultrasound image and ventricular volume curve, marking the position and shape of the left ventricular endocardium, and displaying the type of the above-mentioned section.

For the corresponding relationship of each component or module described above, refer to the description in FIG. 1 above. The image processing module performs the various steps in FIG. 2, and the details are not repeated here. For the specific explanation of each step process, refer to the description of the related content. The image processing module mentioned herein may be constructed by one processor or multiple processors.

For the corresponding relationship of each component or module described above, refer to the description in FIG. 1 above. The image processing module performs the various steps in FIG. 2, and the details are not repeated here. For the specific explanation of each step process, refer to the description of the related content. The image processing module mentioned in this article can use one piece. A processor or a multi-chip processor.

In one embodiment of the present invention, the display prompts the user whether the heartbeat period is incorrect. The system further includes:

An operation control module for receiving a user input control command,

When the display prompts the user that the cardiac cycle is incorrect, the image processing module switches to the manual input mode, and the user can manually input the cardiac cycle by operating the control module.

In one embodiment of the present invention, the display display the identification result of the type of the facet for confirmation by the user, and the system further includes:

An operation control module for receiving a user input control command,

When the control command input by the user confirms the error for the recognition result, the image processing module switches the manual input mode, and modifies and displays the current aspect type according to the control command input by the user via the operation control module.

In one embodiment of the present invention, the display displays the position and shape of the left ventricular endocardium of each frame of the ultrasound image for confirmation by the user, and the system further includes:

An operation control module for receiving a user input control command,

When the control command input by the user confirms the error for the recognition result, the image processing module switches the manual input mode, and modifies and displays the position and shape of the left ventricular endocardium on the ultrasound image according to the control command input by the user via the operation control module.

In one embodiment of the present invention, the display displays the current display and output results for user confirmation and/or correction. The system further includes:

An operation control module for receiving a user input control command,

The image processing module forms an output report based on the confirmation and/or correction command input by the user via the operation control module, and the correction content includes at least: a time corresponding to the end of diastole, a time corresponding to the end of the systole, and a labeled left ventricular endocardium on the ultrasound image. One of the location and shape, and clinical parameters that characterize cardiac function.

In one embodiment of the invention, the display further displays the calculated ventricular volume curve as described above, and/or clinical parameters indicative of cardiac function.

In one of the embodiments of the present invention, the image processing module identifies a cardiac cycle by:

Extracting feature values of each frame image according to the multi-frame ultrasound image continuously acquired over a predetermined time period to generate a characteristic curve;

The characteristic curve is periodically analyzed to identify the cardiac cycle of the target.

In one embodiment of the present invention, the feature value of each frame of the image includes an anatomical measurement value, the characteristic curve being a curve of the anatomical structure measurement value as a function of time; or

The feature value of each frame image is a similarity coefficient, and the feature curve is a similarity coefficient curve.

In one of the embodiments of the present invention, the image processing module identifies a cardiac cycle by:

Select one frame of ultrasound image as the standard frame;

The local extremum is searched in the local area near the time of the standard frame, and the ultrasonic image corresponding to the local extremum is determined; the time corresponding to the standard frame is taken as the start and end points of the time, and a cardiac cycle is obtained.

In one of the embodiments of the present invention, the image processing module identifies the position and shape of the left ventricular endocardium in each frame of ultrasound images within a cardiac cycle by:

Selecting a frame image as a key frame, and performing image segmentation on the left ventricular endocardium of the key frame based on the slice type to obtain a standard endocardial segmentation model;

Based on the standard endocardial segmentation model, the position and shape of the left ventricular endocardium on each frame of ultrasound images within a cardiac cycle is identified.

In one embodiment of the present invention, the image processing module identifies the position of the left ventricular endocardium on each frame of the ultrasound image within a cardiac cycle according to the standard endocardial segmentation model. shape:

According to the standard endocardial segmentation model, the ultrasound images other than the key frame in the cardiac cycle are segmented frame by frame to obtain the position and shape of the left ventricular endocardium on each frame of the ultrasound image; or

Based on the position and shape of the segmented key frame endocardium, the endocardial motion is tracked on other ultrasound images except the key frame in the cardiac cycle to obtain the left ventricular endocardium on each frame of the ultrasound image. Location and shape.

In an embodiment of the present invention, the image processing module automatically selects a frame image as a key frame by performing image segmentation on the left ventricular endocardium of the key frame based on the slice type, and obtains a standard. Endocardial segmentation model:

Edge extraction based on feature points to generate the position and shape of the key frame endocardium; and/or

Using the machine learning method, the position and shape of the key frame endocardium are identified based on the overall model.

In one of the embodiments of the present invention, the image processing module identifies a slice type of the heart in the multi-frame ultrasound image data by:

Mapping one or more frames of ultrasound image data to a feature space, the feature space being constructed by extracting features in the training set image;

Comparing the projection of the ultrasound image in the feature space with the projection of the training image of the known slice type in the feature space, determining the type of the slice of the one or more frames of the ultrasound image.

In one of the embodiments of the present invention, the image processing module processes the ultrasound image in the following manner:

Identifying the location of the interventricular septum in the ultrasound image;

Rotating the ultrasound image according to the position of the chamber interval, so that the long axis direction of the left ventricle in the ultrasound image is vertical;

The ultrasound image is translated to adjust the left ventricular position in the ultrasound image to the center of the image.

The present invention provides a heart volume recognition analysis system and method. The invention utilizes the characteristics of cardiac motion and image processing technology to replace the prior art recognition of the automatic cardiac cycle based on the electrocardiographic signal and the determination of the cardiac motion phase. In the actual use process, there is no need to connect the ECG signal wires and modules, which simplifies the user's workload and improves work efficiency.

A person skilled in the art can understand that all or part of the process of implementing the above embodiment method can be completed by a computer program to instruct related hardware, and the above program can be stored in a computer readable storage medium, and the program is executed. At the time, the flow of the embodiment of each of the above methods may be included. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

The above is a further detailed description of the present invention in connection with the specific preferred embodiments, and the specific embodiments of the present invention are not limited to the description. It will be apparent to those skilled in the art that the present invention may be made without departing from the spirit and scope of the invention.

Claims (34)

1. A method for heart volume recognition analysis, comprising:

   Acquiring a plurality of frames of ultrasound images continuously acquired over a predetermined period of time;

   Identifying a type of face of the heart in the multi-frame ultrasound image data;

   Identify the cardiac cycle;

   Identifying the position and shape of the left ventricular endocardium in each frame of the ultrasound image during a cardiac cycle based on the type of section;

   Calculating a ventricular volume quantitative parameter at a time corresponding to each frame of the ultrasound image according to the position and shape of the left ventricular endocardium, and obtaining a ventricular volume curve;

   The ventricular volume curve is output, and/or clinical parameters characterizing cardiac function are calculated and output based on the ventricular volume curve.
2. A cardiac volume recognition analysis method according to claim 1, wherein the method further comprises:

   Simultaneously displaying an ultrasound image corresponding to the time at which the maximum and minimum values on the ventricular volume curve are located, and marking the position and shape of the left ventricular endocardium; and/or,

   The identified type of facet is displayed.
3. The heart volume identification analysis method according to claim 2, wherein the step of identifying the type of the face of the heart in the multi-frame ultrasound image data further comprises:

   Prompting the user to modify the type of the displayed face;

   The facet type is updated when the user enters the corrected facet type.
4. A cardiac volume recognition analysis method according to claim 1, wherein said step of identifying a position and a shape of a left ventricular endocardium in each frame of an ultrasound image within a cardiac cycle comprises:

   Selecting a frame image as a key frame, and performing image segmentation on the left ventricular endocardium of the key frame based on the slice type to obtain a standard endocardial segmentation model;

   Based on the standard endocardial segmentation model, the position and shape of the left ventricular endocardium on each frame of ultrasound images within a cardiac cycle is identified.
5. A cardiac volume recognition analysis method according to claim 4, wherein the position and shape of the left ventricular endocardium on each frame of the ultrasound image within a cardiac cycle are identified based on said standard endocardial segmentation model The steps include:

   According to the standard endocardial segmentation model, other ultrasounds except for key frames in the cardiac cycle The image is segmented frame by frame to obtain the position and shape of the left ventricular endocardium on each frame of the ultrasound image; or

   Based on the position and shape of the segmented key frame endocardium, the endocardial motion is tracked on other ultrasound images except the key frame in the cardiac cycle to obtain the left ventricular endocardium on each frame of the ultrasound image. Location and shape.
6. A heart volume recognition analysis method according to claim 4, wherein said automatically selecting a frame image as a key frame, and performing image segmentation on the left ventricular endocardium of said key frame based on said slice type, obtaining a standard The steps of the endocardial segmentation model include:

   Edge extraction based on feature points to generate the position and shape of the key frame endocardium; and/or

   Using the machine learning method, the position and shape of the key frame endocardium are identified based on the overall model.
7. A cardiac volume recognition analysis method according to claim 1, wherein said step of identifying a type of face of a heart in said multi-frame ultrasound image data comprises:

   A feature of one or more frames of the ultrasound image in the multi-frame ultrasound image is compared with a feature of a training image of a known slice type to obtain a slice type of the multi-frame ultrasound image.
8. A cardiac volume recognition analysis method according to claim 1, wherein said step of identifying a type of face of a heart in said multi-frame ultrasound image data comprises:

   Mapping one or more frames of ultrasound images to a feature space, the feature space being constructed by extracting features in the training set image;

   Comparing the projection of the ultrasound image in the feature space with the projection of the training image of the known slice type in the feature space, determining the type of the slice of the one or more frames of the ultrasound image.
9. A cardiac volume recognition analysis method according to claim 7, wherein said comparing the feature of one or more frames of the ultrasound image in said multi-frame ultrasound image with the characteristics of a training image of a known slice type Before the steps include:

   Identifying the location of the interventricular septum in the ultrasound image;

   Rotating the ultrasound image according to the position of the chamber interval, so that the long axis direction of the left ventricle in the ultrasound image is vertical;

   The ultrasound image is translated to adjust the left ventricular position in the ultrasound image to the center of the image.
10. The cardiac volume recognition analysis method according to claim 7, wherein the determining the type of the slice of the one or more frames of the ultrasound image further comprises:

    Comparing a slice type of one frame of the ultrasound image in the multi-frame ultrasound image to a slice type corresponding to the multi-frame ultrasound image; or

    Determining a type of the slice corresponding to the multi-frame ultrasound image based on a type of the slice of each frame of the ultrasound image in the multi-frame ultrasound image.
11. A cardiac volume recognition analysis method according to claim 1, wherein said step of identifying a cardiac cycle comprises:

    Extracting feature values of each frame image according to the multi-frame ultrasound image continuously acquired over a predetermined time period to generate a characteristic curve;

    Periodic analysis of the characteristic curve identifies a cardiac cycle.
12. The cardiac volume recognition analysis method according to claim 11, wherein the feature value of each frame of the image comprises an anatomical structure measurement value, wherein the characteristic curve is a curve of an anatomical structure measurement value as a function of time; or

    The feature value of each frame image is a similarity coefficient, and the feature curve is a similarity coefficient curve.
13. A cardiac volume recognition analysis method according to claim 1, wherein the step of identifying a cardiac cycle comprises:

    Select one frame of ultrasound image as the standard frame;

    The local extremum is searched in the local area near the time of the standard frame, and the ultrasonic image corresponding to the local extremum is determined; the time corresponding to the standard frame is taken as the start and end points of the time, and a cardiac cycle is obtained.
14. A cardiac volume recognition analysis method according to claim 1, wherein the method further comprises:

    The calculated ventricular volume curve is displayed, and/or clinical parameters characterizing cardiac function.
15. The cardiac volume recognition analysis method according to claim 1, wherein the step of identifying the cardiac cycle further comprises:

    After the failure of the cardiac cycle recognition, it is switched to the manual input mode for obtaining the cardiac cycle manually input by the user.
16. A cardiac volume recognition analysis method according to claim 4, wherein said step of identifying the position and shape of the left ventricular endocardium in each frame of the ultrasound image within one cardiac cycle comprises:

    Displaying the key frame;

    Prompting the user to manually intervene in the image segmentation process of the left ventricular endocardium in the key frame;

    The standard endocardial segmentation is obtained based on an adjustment or input result of the user on the key frame model.
17. A cardiac volume recognition analysis method according to claim 1 or 4, wherein said step of identifying a position and a shape of a left ventricular endocardium in each frame of an ultrasound image within a cardiac cycle, or said After the step of the image segmentation of the left ventricular endocardium of the key frame to obtain a standard endocardial segmentation model, the method further includes:

    Determining the segmentation result of the position and shape of the left ventricular endocardium, when it is determined that the segmentation result is incorrect, jumps to the manual input mode to prompt the user to input the position and shape of the left ventricular endocardium on the ultrasound image.
18. The cardiac volume recognition analysis method according to claim 2, wherein the method further comprises:

    Prompt the user to confirm and/or correct the current display and output results;

    Based on the user's confirmation and/or correction, an output report is formed, which includes at least: a time corresponding to the end of diastole, a time corresponding to the end of systole, a position and shape of the left ventricular endocardium marked on the ultrasound image, and a function of characterizing the heart. One of the clinical parameters.
19. A cardiac volume recognition and analysis system, comprising:

    An ultrasound image acquisition module, configured to acquire a multi-frame ultrasound image continuously acquired over a predetermined time period;

    a facet type identification module, configured to identify a facet type of the heart in the multi-frame ultrasound image data;

    a cardiac cycle recognition module for identifying a cardiac cycle;

    a contour obtaining module for identifying a position and a shape of a left ventricular endocardium in each frame of the ultrasound image within a cardiac cycle based on the type of the slice;

    a ventricular volume curve generating module, configured to calculate a ventricular volume quantitative parameter at a time corresponding to each frame of the ultrasound image according to the position and shape of the left ventricular endocardium, and obtain a ventricular volume curve;

    A clinical parameter output module for outputting a ventricular volume curve and/or calculating and outputting clinical parameters indicative of cardiac function from the ventricular volume curve.
20. A cardiac volume recognition analysis system according to claim 19, wherein said system further comprises:

    Displaying a marking module for displaying an ultrasound image corresponding to a time at which the maximum and minimum values on the ventricular volume curve are located, and marking the position and shape of the left ventricular endocardium; and/or,

    A facet type display module for displaying the identified type of facet.
21. A cardiac volume recognition and analysis system, comprising:

    Probe

    a transmitting circuit for transmitting an ultrasonic beam to a heart region of the target object;

    a receiving circuit and a beam combining module for obtaining an ultrasonic echo signal;

    An image processing module, configured to obtain, according to the ultrasonic echo signal, a multi-frame ultrasound image continuously acquired over a predetermined time period, identify a type of the heart surface in the ultrasound image, identify a cardiac cycle, and identify each frame of the ultrasound image in a cardiac cycle The position and shape of the middle left ventricle endocardium, and calculate the ventricular volume quantitative parameter at the time corresponding to each frame of the ultrasound image to obtain a ventricular volume curve;

    a display for displaying the ultrasound image, marking the position and shape of the left ventricular endocardium, and displaying the type of section.
22. A heart volume recognition and analysis system according to claim 21, wherein said display prompts the user whether the cardiac cycle is recognized as being incorrect, and the system further comprises:

    An operation control module for receiving a user input control command;

    When the display prompts the user that the cardiac cycle is incorrect, the image processing module switches to the manual input mode, and the user can manually input the cardiac cycle by operating the control module.
23. A cardiac volume recognition and analysis system according to claim 21, wherein said display displays said recognition result of said aspect type for confirmation by a user, said system further comprising:

    An operation control module for receiving a user input control command;

    When the control command input by the user confirms the error for the recognition result, the image processing module switches the manual input mode, and modifies and displays the current aspect type according to the control command input by the user via the operation control module.
24. A cardiac volume recognition analysis system according to claim 21, wherein said display displays the position and shape of the left ventricular endocardium of each frame of the ultrasound image for confirmation by the user, said system further comprising:

    An operation control module for receiving a user input control command;

    When the control command input by the user confirms the error for the recognition result, the image processing module switches the manual input mode, and modifies and displays the position and shape of the left ventricular endocardium on the ultrasound image according to the control command input by the user via the operation control module.
25. A cardiac volume recognition analysis system according to claim 21, wherein said display displays current display and output results for user confirmation and/or correction, said system further comprising:

    An operation control module for receiving a user input control command;

    The image processing module forms an output report based on the confirmation and/or correction command input by the user through the operation control module, and the correction content includes at least: a time corresponding to the end of diastole, a time corresponding to the end of the contraction, One of the location and shape of the left ventricular endocardium labeled on the ultrasound image, and the clinical parameters characterizing cardiac function.
26. A cardiac volume recognition analysis system according to claim 21, wherein said display further displays said calculated ventricular volume curve and/or clinical parameters indicative of cardiac function.
27. A cardiac volume recognition analysis system according to claim 21, wherein said image processing module identifies a cardiac cycle by:

    Extracting feature values of each frame image according to the multi-frame ultrasound image continuously acquired over a predetermined time period to generate a characteristic curve;

    The characteristic curve is periodically analyzed to identify the cardiac cycle of the target.
28. A cardiac volume recognition analysis system according to claim 27, wherein said characteristic value of each frame of image comprises an anatomical measurement value, said characteristic curve being a curve of anatomical measurement value as a function of time;

    The feature value of each frame image is a similarity coefficient, and the feature curve is a similarity coefficient curve.
29. A cardiac volume recognition analysis system according to claim 21, wherein said image processing module identifies a cardiac cycle by:

    Select one frame of ultrasound image as the standard frame;

    The local extremum is searched in the local area near the time of the standard frame, and the ultrasonic image corresponding to the local extremum is determined; the time corresponding to the standard frame is taken as the start and end points of the time, and a cardiac cycle is obtained.
30. A cardiac volume recognition analysis system according to claim 21, wherein said image processing module identifies the position and shape of the left ventricular endocardium in each frame of the ultrasound image in a cardiac cycle by:

    Selecting a frame image as a key frame, and performing image segmentation on the left ventricular endocardium of the key frame based on the slice type to obtain a standard endocardial segmentation model;

    Based on the standard endocardial segmentation model, the position and shape of the left ventricular endocardium on each frame of ultrasound images within a cardiac cycle is identified.
31. A cardiac volume recognition analysis system according to claim 30, wherein said image processing module identifies a left image of each frame within a cardiac cycle according to said standard endocardial segmentation model in the following manner Location and shape of the ventricular endocardium:

    According to the standard endocardial segmentation model, other ultrasounds except for key frames in the cardiac cycle The image is segmented frame by frame to obtain the position and shape of the left ventricular endocardium on each frame of the ultrasound image; or

    Based on the position and shape of the segmented key frame endocardium, the endocardial motion is tracked on other ultrasound images except the key frame in the cardiac cycle to obtain the left ventricular endocardium on each frame of the ultrasound image. Location and shape.
32. 30. A cardiac volume recognition analysis system according to claim 30, wherein said image processing module automatically selects a frame image as a key frame by, based on said slice type, a left ventricle of said key frame The film was image segmented to obtain a standard endocardial segmentation model:

    Edge extraction based on feature points to generate the position and shape of the key frame endocardium; and/or

    Using the machine learning method, the position and shape of the key frame endocardium are identified based on the overall model.
33. A cardiac volume recognition analysis system according to claim 21, wherein said image processing module identifies a slice type of a heart in said multi-frame ultrasound image data by:

    A feature of one or more frames of the ultrasound image in the multi-frame ultrasound image is compared with a feature of a training image of a known slice type to obtain a slice type of the multi-frame ultrasound image.
34. A cardiac volume recognition analysis system according to claim 33, wherein said image processing module performs a feature of one or more frames of the ultrasound image in said multi-frame ultrasound image with a training image of a known slice type Before the characteristics are compared, the following processing is also performed:

    Identifying the location of the interventricular septum in the ultrasound image;

    Rotating the ultrasound image according to the position of the chamber interval, so that the long axis direction of the left ventricle in the ultrasound image is vertical;

    The ultrasound image is translated to adjust the left ventricular position in the ultrasound image to the center of the image.

Images