CN118750024A - A method for assisting detection of cardiovascular conditions in internal medicine - Google Patents

Abstract

The present invention relates to the field of image data processing technology, and in particular to a cardiovascular condition auxiliary detection method for internal medicine, the method comprising: obtaining a target ultrasound cardiogram of a patient to be detected at each preset time within a preset number of heartbeat cycles; screening out the cardiac structure contour from each heart section image, and determining a clear identification factor to be enhanced corresponding to each heart section image; determining a low resolution factor corresponding to each heartbeat cycle; determining a target standard deviation corresponding to each frame of the target ultrasound cardiogram according to the clear identification factor to be enhanced and the low resolution factor; and upsampling each frame of the target ultrasound cardiogram through a Gaussian interpolation kernel according to the target standard deviation corresponding to each frame of the target ultrasound cardiogram to obtain a target image, wherein the target standard deviation is the standard deviation obeyed by the Gaussian interpolation kernel. The present invention improves the effect of upsampling the ultrasound cardiogram by performing image data processing on the target ultrasound cardiogram.

Description

A method for assisting detection of cardiovascular conditions in internal medicine

Technical Field

The present invention relates to the technical field of image data processing, and in particular to an auxiliary detection method for cardiovascular conditions in internal medicine.

Background Art

With the development of science and technology, the application of image data processing is becoming more and more extensive. For example, it can be used in the auxiliary detection of cardiovascular conditions in internal medicine. Specifically, doctors can make a preliminary diagnosis of patients by observing the collected echocardiograms. Due to the influence of environmental factors and other factors, the resolution of the collected echocardiograms may be low, which makes it difficult to display some important details in the echocardiograms. Therefore, it is often necessary to upsample the collected echocardiograms to improve the resolution of the echocardiograms. At present, when upsampling an image, the method usually used is: preset a standard deviation for the Gaussian interpolation kernel, and upsample the image through the Gaussian interpolation kernel. Among them, the standard deviation preset for the Gaussian interpolation kernel is often set based on manual experience.

However, when a standard deviation is preset for the Gaussian interpolation kernel and multiple acquired echocardiogram frames are upsampled using the Gaussian interpolation kernel, the following technical problems often occur:

Due to the influence of environmental factors, the resolutions of different frames of echocardiograms are often different. Therefore, the standard deviations required for upsampling different frames of echocardiograms are often different. Therefore, if only one standard deviation is preset, it may lead to poor upsampling effects on some echocardiograms. Secondly, since the standard deviation preset for the Gaussian interpolation kernel is often set based on manual experience, the setting results are often affected by human subjective factors, so the setting results are often inaccurate, resulting in poor upsampling effects on echocardiograms.

Summary of the invention

In order to solve the technical problem of poor upsampling effect of echocardiogram, the present invention proposes an auxiliary detection method for cardiovascular condition in internal medicine.

In a first aspect, the present invention provides a method for auxiliary detection of cardiovascular conditions for internal medicine, the method comprising:

Acquire a target echocardiogram of the patient to be tested at each preset time within a preset number of heartbeat cycles, wherein the target echocardiogram includes a heart section image;

Screening out the cardiac structure contour from each cardiac section image, and determining the identification clarity enhancement factor corresponding to each cardiac section image according to the number of edge lines on the cardiac structure contour in each cardiac section image and the grayscale difference between the pixels on the cardiac structure contour;

Determine a low-resolution factor corresponding to each heartbeat cycle according to the grayscale difference between each heartbeat cycle and the target echocardiogram in other heartbeat cycles, and the corresponding physical sign information of the patient to be detected;

Determine the target standard deviation corresponding to each frame of the target echocardiogram according to the identification clear enhancement factor corresponding to the heart section image contained in each frame of the target echocardiogram and the low resolution factor corresponding to the heartbeat cycle to which each frame of the target echocardiogram belongs;

According to the target standard deviation corresponding to each frame of the target echocardiogram, each frame of the target echocardiogram is upsampled by a Gaussian interpolation kernel to obtain a target image, wherein the target standard deviation is the standard deviation obeyed by the Gaussian interpolation kernel.

In combination with the first aspect, in a possible implementation, determining the identification clarity enhancement factor corresponding to each cardiac section image according to the number of edge lines on the cardiac structure contour in each cardiac section image and the grayscale difference between pixels on the cardiac structure contour includes:

The absolute value of the difference between the grayscale values corresponding to every two pixel points on the contour of each cardiac structure is determined as the reference grayscale difference between the two pixel points;

According to the reference grayscale difference between all pixel points on each cardiac structure contour, the grayscale representative difference corresponding to each cardiac structure contour is determined, wherein the reference grayscale difference and the grayscale representative difference are positively correlated;

Perform edge detection on each cardiac structure contour to obtain edge lines, and determine the edge identification index to be adjusted corresponding to each cardiac structure contour according to the number of edge lines on each cardiac structure contour and the corresponding grayscale representative difference, wherein the number of edge lines on the cardiac structure contour and the corresponding grayscale representative difference are both positively correlated with the corresponding edge identification index to be adjusted;

According to the edge recognition indexes to be adjusted corresponding to all cardiac structure contours in each cardiac section image and the grayscale values corresponding to the pixels on each cardiac structure contour in each cardiac section image, the recognition clarity factor to be enhanced corresponding to each cardiac section image is determined.

In combination with the first aspect, in a possible implementation, determining the identification clarity factor to be enhanced corresponding to each cardiac section image according to the edge identification indicators to be adjusted corresponding to all cardiac structure contours in each cardiac section image and the grayscale values corresponding to the pixels on each cardiac structure contour in each cardiac section image includes:

The union of the preset neighborhoods corresponding to all the pixel points on each cardiac structure contour is determined as the reference neighborhood corresponding to each cardiac structure contour;

Determine the contour index to be adjusted corresponding to each cardiac structure contour according to each cardiac structure contour and its corresponding reference neighborhood and edge identification index to be adjusted;

According to the contour adjustment index corresponding to all cardiac structure contours in each cardiac section image, the identification clarity enhancement factor corresponding to each cardiac section image is determined, wherein the contour adjustment index and the identification clarity enhancement factor are positively correlated.

In combination with the first aspect, in a possible implementation, determining the contour index to be adjusted corresponding to each cardiac structure contour according to each cardiac structure contour and its corresponding reference neighborhood and edge identification index to be adjusted includes:

The mean of the grayscale values corresponding to all the pixels on each cardiac structure contour is determined as the first grayscale representative factor corresponding to each cardiac structure contour;

The average of the grayscale values corresponding to all the pixels in the reference neighborhood corresponding to each cardiac structure contour is determined as the second grayscale representative factor corresponding to each cardiac structure contour;

Determine the difference between the first grayscale representative factor and the second grayscale representative factor corresponding to each cardiac structure contour as the grayscale contrast difference corresponding to each cardiac structure contour;

According to the edge identification index to be adjusted and the grayscale contrast difference corresponding to each cardiac structure contour, the contour index to be adjusted corresponding to each cardiac structure contour is determined, wherein the edge identification index to be adjusted is positively correlated with the contour index to be adjusted, and the grayscale contrast difference is negatively correlated with the contour index to be adjusted.

In combination with the first aspect above, in a possible implementation, determining the low-resolution factor corresponding to each heartbeat cycle according to the grayscale difference between each heartbeat cycle and the target echocardiogram in other heartbeat cycles, and the physical sign information corresponding to the patient to be detected, includes:

According to the grayscale difference between the target echocardiograms within each two heartbeat cycles, the information change amount between each two heartbeat cycles is determined;

Determining a resolution evaluation value corresponding to the patient to be detected according to the physical sign information corresponding to the patient to be detected;

The low resolution factor corresponding to each heartbeat cycle is determined according to the resolution evaluation value and the information change amount between each heartbeat cycle and other heartbeat cycles, wherein the information change amount and the resolution evaluation value are both positively correlated with the low resolution factor.

In combination with the first aspect above, in a possible implementation, determining the amount of information change between every two heartbeat cycles according to the grayscale difference between the target echocardiograms within every two heartbeat cycles includes:

Determine any one heartbeat cycle as a marking cycle, determine any one heartbeat cycle outside the marking cycle as a reference cycle, determine any one frame of target echocardiogram within the marking cycle as a marking image, and select a target echocardiogram whose corresponding frame number is the same as that of the marking image from within the reference cycle as a reference image;

Determine the absolute value of the difference between the average grayscale corresponding to the marked image and the average grayscale corresponding to the reference image as the target grayscale difference between the marked image and the reference image;

Determine the absolute value of the difference between the grayscale variance corresponding to the marked image and the grayscale variance corresponding to the reference image as the grayscale distribution difference between the marked image and the reference image;

Determining a target difference factor between the marked image and the reference image according to a target grayscale difference between the marked image and the reference image, and a grayscale distribution difference between the marked image and the reference image, wherein both the target grayscale difference and the grayscale distribution difference are positively correlated with the target difference factor;

The information variation between every two heartbeat cycles is determined according to all target difference factors within every two heartbeat cycles, wherein the target difference factor is positively correlated with the information variation.

In combination with the first aspect above, in a possible implementation manner, determining the resolution evaluation value corresponding to the patient to be detected according to the physical sign information corresponding to the patient to be detected includes:

Inputting the vital sign information corresponding to the patient to be detected into a pre-trained resolution evaluation network, and outputting the resolution evaluation value corresponding to the patient to be detected through the pre-trained resolution evaluation network, wherein the training process of the resolution evaluation network includes:

Construct a resolution assessment network;

Acquire a sample vital sign information set, wherein the resolution evaluation value corresponding to the sample vital sign information in the sample vital sign information set is known;

The sample vital sign information set is used as a training set, and the resolution assessment value corresponding to the sample vital sign information is used as a training label to train the constructed resolution assessment network to obtain a trained resolution assessment network.

In combination with the first aspect, in a possible implementation, determining the target standard deviation corresponding to each frame of the target echocardiogram according to the identification clarity factor to be enhanced corresponding to the cardiac section image contained in each frame of the target echocardiogram and the low resolution factor corresponding to the heart cycle to which each frame of the target echocardiogram belongs includes:

The Euclidean norm between the identification clear factor to be enhanced corresponding to the cardiac section image contained in each frame of the target echocardiogram and the low resolution factor corresponding to the heart cycle to which each frame of the target echocardiogram belongs is determined as the target standard deviation corresponding to each frame of the target echocardiogram.

In combination with the first aspect above, in a possible implementation manner, screening out the cardiac structure contour from each cardiac section image includes:

Each cardiac section image is input into a pre-trained cardiac structure contour recognition network, and the pre-trained cardiac structure contour recognition network is used to recognize the cardiac structure contour in each cardiac section image.

In combination with the first aspect above, in a possible implementation, the training process of the cardiac structure contour recognition network includes:

Construct a cardiac structure contour recognition network;

Acquire a set of sample cardiac section images, wherein the cardiac structure contours are marked in the sample cardiac section images in the set of sample cardiac section images;

The constructed cardiac structure contour recognition network is trained using the sample cardiac section image set as a training set and the cardiac structure contours annotated in the sample cardiac section images as training labels to obtain a trained cardiac structure contour recognition network.

The present invention has the following beneficial effects:

The present invention is an auxiliary detection method for cardiovascular conditions in internal medicine, which solves the technical problem of poor upsampling effect of the echocardiogram by processing the image data of the target echocardiogram, and improves the effect of upsampling the echocardiogram. The present invention comprehensively considers multiple indicators related to resolution, such as the identification of clear factors to be enhanced and low resolution factors, and quantifies the target standard deviation required when upsampling each frame of the target echocardiogram, so that adaptive upsampling of each frame of the target echocardiogram can be achieved, thereby improving the upsampling effect. Secondly, the process of quantifying the target standard deviation of the present invention is relatively objective, which reduces the influence of human subjective factors to a certain extent, thereby improving the accuracy of determining the target standard deviation, and then improving the effect of upsampling each frame of the target echocardiogram.

In a second aspect, the present invention provides a cardiovascular condition auxiliary detection system for internal medicine, the system comprising:

An image acquisition module, used to acquire a target echocardiogram of the patient to be tested at each preset time within a preset number of heartbeat cycles, wherein the target echocardiogram includes a heart section image;

A screening and determination module is used to screen out the cardiac structure contour from each cardiac section image, and determine the identification clarity enhancement factor corresponding to each cardiac section image according to the number of edge lines on the cardiac structure contour in each cardiac section image and the grayscale difference between the pixels on the cardiac structure contour;

A resolution factor determination module, used to determine a low resolution factor corresponding to each heartbeat cycle according to the grayscale difference between each heartbeat cycle and the target echocardiogram in other heartbeat cycles, and the physical sign information corresponding to the patient to be detected;

A standard deviation determination module is used to determine a target standard deviation corresponding to each frame of the target echocardiogram according to a clear identification factor to be enhanced corresponding to the cardiac section image contained in each frame of the target echocardiogram and a low resolution factor corresponding to the heart cycle to which each frame of the target echocardiogram belongs;

The upsampling module is used to upsample each frame of the target echocardiogram through a Gaussian interpolation kernel according to a target standard deviation corresponding to each frame of the target echocardiogram to obtain a target image, wherein the target standard deviation is the standard deviation obeyed by the Gaussian interpolation kernel.

In a third aspect, a server is provided, comprising a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes the method in the first aspect or any possible implementation of the first aspect.

In a fourth aspect, a computer program product is provided, comprising: a computer program code, which, when executed on a computer, enables the computer to execute the method in the first aspect or any possible implementation of the first aspect.

In a fifth aspect, a computer-readable storage medium is provided, which stores a computer program code. When the computer program code runs on a computer, the computer executes the method in the above-mentioned first aspect or any possible implementation of the first aspect.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

FIG1 is a flow chart of a cardiovascular condition auxiliary detection method for internal medicine of the present invention;

FIG2 is a flow chart of a method for determining a clear identification factor to be enhanced according to the present invention;

FIG3 is another flow chart of a method for determining a clear factor to be enhanced according to the present invention;

FIG4 is a flow chart of a method for determining a contour index to be adjusted according to the present invention;

FIG5 is a flow chart of a method for determining a low resolution factor according to the present invention;

FIG6 is a schematic diagram of the composition structure of a cardiovascular condition auxiliary detection system for internal medicine according to the present invention;

FIG. 7 is a schematic diagram of the structure of a computer device according to the present invention.

DETAILED DESCRIPTION

In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the technical solutions proposed by the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

Referring to FIG1 , a flow chart of some embodiments of a cardiovascular condition auxiliary detection method for internal medicine according to the present invention is shown. The cardiovascular condition auxiliary detection method for internal medicine comprises the following steps:

Step S1, obtaining a target echocardiogram of a patient to be tested at each preset time within a preset number of heartbeat cycles.

Among them, the patient to be detected may be a patient to be tested for cardiovascular condition. The preset number may be a preset number. For example, the preset number may be 6. A heartbeat cycle represents a heartbeat process of the patient to be detected. For example, the duration from the start of a heartbeat of the patient to be detected to the start of the next heartbeat of the patient to be detected may be regarded as a heartbeat cycle. The end time of the previous heartbeat cycle may be the start time of the next heartbeat cycle. The preset time may be a preset time. The duration between adjacent preset times may be the same. For example, the duration between adjacent preset times may be 0.15 seconds. The target echocardiogram may be an echocardiogram related to the heart, and the target echocardiogram in the embodiment of the present invention may be a grayscale image. The echocardiogram may represent the structural changes, blood flow velocity and direction, and other states during heart contraction and diastole. For example, the echocardiogram may include a heart section image, tissue Doppler imaging, and a strain image. Therefore, the target echocardiogram may include a heart section image, a tissue Doppler imaging, and a strain image. The heart section image may be a section image related to the patient's heart. For example, a cross-sectional image of the heart may include cardiac structures such as the atria, ventricles, valves, and great vessels.

It should be noted that echocardiography is a non-invasive examination method that uses high-frequency sound waves to generate cardiac images. It can provide detailed information on the structure and function of the heart and is very important for the diagnosis and treatment of cardiovascular diseases. Among them, high-frequency sound waves are also called ultrasound. The principle of obtaining echocardiography is as follows: the echocardiography instrument emits high-frequency sound waves, which penetrate the skin and body tissues to reach the heart. When the sound waves encounter the structure of the heart, they are reflected back to the instrument, and the receiver captures these echoes. Then, the instrument converts these data into real-time cardiac images. Secondly, when the resolution of the echocardiogram is insufficient, it is often necessary to upsample the echocardiogram to make the differences between the cardiac structures more obvious and to make the speed and direction of blood flow more clearly visible, so as to help doctors more easily identify the boundaries between different tissues and structures, as well as to evaluate cardiac hemodynamics and detect valve function. In general, the duration corresponding to different heartbeat cycles of the same patient in a short period of time is often the same. In the process of auxiliary detection of cardiovascular conditions of patients, echocardiograms are often not collected for a long time. Therefore, the duration corresponding to the above-mentioned preset number of heartbeat cycles is often the same.

As an example, an ultrasonic probe included in an ultrasonic cardiography instrument may be used to perform ultrasonic scanning to acquire an ultrasonic cardiogram of the patient to be detected at each preset time as a target ultrasonic cardiogram.

It should be noted that, generally speaking, in the process of obtaining an echocardiogram, the patient often needs to be placed in a comfortable position. In order to facilitate the sliding of the probe and the conduction of ultrasonic waves, it is often necessary to apply an appropriate amount of gel on the patient's chest; according to the patient's body shape, age and the part to be examined, an appropriate ultrasound probe is selected. Commonly used ultrasound probes include linear probes and convex probes; the ultrasound probe is placed on the patient's chest, usually in the precordial area on the left side of the sternum, and scans are performed at different positions as needed, including the left edge of the sternum, the apex area, etc.; the position and angle of the probe are adjusted to observe the heart structure in the best way. It should be noted that when acquiring each target echocardiogram, the position and angle of the probe remain unchanged. At this time, the position and angle of the probe can be set based on manual experience; then start the ultrasonic scan, and observe the real-time display of the echocardiogram at the same time; during the scan, the structural changes, blood flow velocity and direction during heart contraction and relaxation, and other information images are recorded to form an echocardiogram.

Step S2, screening out the cardiac structure contour from each cardiac section image, and determining the identification clarity enhancement factor corresponding to each cardiac section image according to the number of edge lines on the cardiac structure contour in each cardiac section image and the grayscale difference between the pixels on the cardiac structure contour.

The cardiac structure contour may represent the boundary of the cardiac structure. It should be noted that due to the influence of various factors such as the environment, the boundary of the cardiac structure may not be completely clearly displayed, so the boundary of the cardiac structure may be composed of multiple edge lines. Therefore, the cardiac structure contour may include multiple edge lines.

As an example, a method for filtering out the cardiac structure contour from each cardiac section image may be: performing superpixel segmentation on each cardiac section image, treating each obtained superpixel block as a cardiac structure, and determining the boundary of each superpixel block as the cardiac structure contour.

Optionally, the method for filtering out the cardiac structure contour from each cardiac section image may also be: inputting each cardiac section image into a pre-trained cardiac structure contour recognition network, and identifying the cardiac structure contour in each cardiac section image through the pre-trained cardiac structure contour recognition network.

The heart structure contour recognition network may be used to recognize the heart structure contour. For example, the heart structure contour recognition network may be a CNN (Convolutional Neural Networks).

For example, the training process of a cardiac structure contour recognition network includes the following steps:

The first step is to build a cardiac structure contour recognition network.

For example, a CNN can be constructed as a pre-trained cardiac structure contour recognition network.

The second step is to obtain a set of sample heart section images.

The sample cardiac section images in the sample cardiac section image set are annotated with cardiac structure contours.

The third step is to use the above-mentioned sample cardiac section image set as a training set and the cardiac structure contours annotated in the sample cardiac section images as training labels to train the constructed cardiac structure contour recognition network to obtain a trained cardiac structure contour recognition network.

As another example, according to the number of edge lines on the contour of the cardiac structure in each cardiac section image and the grayscale difference between the pixels on the contour of the cardiac structure, the process of determining the clear identification factor to be enhanced corresponding to each cardiac section image can be shown in FIG. 2, and can specifically include the following steps:

Step 201: determine the absolute value of the difference between the grayscale values corresponding to every two pixel points on each cardiac structure contour as the reference grayscale difference between the two pixel points.

Step 202: Determine the grayscale representative difference corresponding to each cardiac structure contour according to the reference grayscale difference between all pixel points on each cardiac structure contour.

Among them, the reference grayscale difference can be positively correlated with the grayscale representative difference.

Step 203 , edge detection is performed on each cardiac structure contour to obtain edge lines, and the edge recognition index to be adjusted corresponding to each cardiac structure contour is determined according to the number of edge lines on each cardiac structure contour and the corresponding grayscale representative difference.

Among them, the number of edge lines on the contour of the heart structure and the corresponding grayscale representative differences can be positively correlated with the corresponding edge recognition index to be adjusted.

For example, the formula for determining the edge recognition index to be adjusted corresponding to the cardiac structure contour may be:

Wherein, JV _r is the edge recognition index to be adjusted corresponding to the r-th cardiac structure contour in the cardiac section image. _{r is the serial number of the cardiac structure contour in the cardiac section image. G r} is the number of edge lines on the r-th cardiac structure contour in the cardiac section image. _{n r} is the number of pixels on the r-th cardiac structure contour in the cardiac section image. i and j are the serial numbers of different pixels on the r-th cardiac structure contour in the cardiac section image. || is the absolute value function. _{g ri} is the grayscale value corresponding to the i-th pixel on the r-th cardiac structure contour in the cardiac section image. _{g rj} is the grayscale value corresponding to the j-th pixel on the r-th cardiac structure contour in the cardiac section image. |g _ri -g _rj | is the reference grayscale difference between the i-th pixel and the j-th pixel. It is the grayscale representation difference corresponding to the rth cardiac structure contour in the cardiac section image.

It should be noted that when G _r is larger, it often means that the edge line on the r-th cardiac structure contour is more discontinuous, the recognizability of the r-th cardiac structure contour is lower, and the resolution of the cardiac section image to which the r-th cardiac structure contour belongs may be lower. When JV r is larger, it often indicates that the grayscale distribution on the r-th cardiac structure contour is more uneven, the recognizability of the r-th cardiac structure contour is lower, and the resolution of the cardiac section image to which the r-th cardiac structure contour belongs may be lower. Therefore, when JV _r is larger, it often indicates that the edge line on the r-th cardiac structure contour is more discontinuous, and the grayscale distribution on the r-th cardiac structure contour is more uneven; it often indicates that the recognizability of the r-th cardiac structure contour is lower, and the resolution of the cardiac section image to which the r-th cardiac structure contour belongs may be lower.

Step 204, determining the identification clarity enhancement factor corresponding to each cardiac section image according to the edge identification adjustment index corresponding to all cardiac structure contours in each cardiac section image and the grayscale value corresponding to the pixel points on each cardiac structure contour in each cardiac section image.

For example, according to the edge recognition indexes to be adjusted corresponding to all cardiac structure contours in each cardiac section image, and the grayscale values corresponding to the pixels on the contours of each cardiac structure in each cardiac section image, the process of determining the recognition clarity factor to be enhanced corresponding to each cardiac section image can be shown in FIG. 3, and can specifically include the following steps:

Step 301: determine the union of preset neighborhoods corresponding to all pixel points on each cardiac structure contour as a reference neighborhood corresponding to each cardiac structure contour.

The preset neighborhood may be a preset neighborhood, for example, a 5×5 neighborhood.

Step 302: Determine the contour index to be adjusted corresponding to each cardiac structure contour according to each cardiac structure contour and its corresponding reference neighborhood and edge identification index to be adjusted.

For example, the process of the method for determining the index to be adjusted of the profile may be as shown in FIG4 , and may specifically include the following steps:

Step 401: determine the mean value of the grayscale values corresponding to all the pixels on each cardiac structure contour as the first grayscale representative factor corresponding to each cardiac structure contour.

Step 402: Determine the average of the grayscale values corresponding to all pixels in the reference neighborhood corresponding to each cardiac structure contour as a second grayscale representative factor corresponding to each cardiac structure contour.

Step 403: Determine the difference between the first grayscale representative factor and the second grayscale representative factor corresponding to each cardiac structure contour as the grayscale contrast difference corresponding to each cardiac structure contour.

For example, the difference between the first grayscale representative factor and the second grayscale representative factor corresponding to each cardiac structure contour may be determined as the grayscale contrast difference corresponding to each cardiac structure contour.

For another example, the absolute value of the difference between the first grayscale representative factor and the second grayscale representative factor corresponding to each cardiac structure contour may be determined as the grayscale contrast difference corresponding to each cardiac structure contour.

It should be noted that, generally speaking, the outline of the heart structure is often brighter than its surrounding area, so the grayscale of the outline of the heart structure is often higher than the grayscale of the surrounding area. Therefore, when calculating the difference between the two, the difference can be calculated by difference calculation or by the absolute value of the difference. At this time, the difference between the two is the grayscale contrast difference.

Step 404 , determining the contour index to be adjusted corresponding to each cardiac structure contour according to the edge recognition index to be adjusted and the grayscale contrast difference corresponding to each cardiac structure contour.

The edge identification index to be adjusted may be positively correlated with the contour index to be adjusted, and the grayscale contrast difference may be negatively correlated with the contour index to be adjusted.

Step 303 , determining a recognition clarity factor to be enhanced corresponding to each cardiac section image according to the contour adjustment indicators corresponding to all cardiac structure contours in each cardiac section image.

Among them, the contour index to be adjusted can be positively correlated with the identification clarity factor to be enhanced.

For example, the formula for determining the clear identification factor to be enhanced corresponding to the cardiac section image can be:

Wherein, E _a is the identification clarity enhancement factor corresponding to the a-th cardiac section image. _a is the serial number of the cardiac section image. Ra is the number of cardiac structure contours in the a-th cardiac section image. r is the serial number of the cardiac structure contour in the cardiac section image. _{JV ar} is the edge identification adjustment index corresponding to the r-th cardiac structure contour in the a-th cardiac section image. _{h ar1} is the first grayscale representative factor corresponding to the r-th cardiac structure contour in the a-th cardiac section image. _{h ar2} is the second grayscale representative factor corresponding to the r-th cardiac structure contour in the a-th cardiac section image. It is a pre-set factor greater than 0, mainly used to prevent the denominator from being 0, such as, It can be 0.001. _{h ar1} -h _ar2 can represent the grayscale contrast difference corresponding to the r-th cardiac structure contour. The contour index to be adjusted corresponding to the r-th cardiac structure contour can be represented.

It should be noted that when JV _ar is larger, it often indicates that the recognizability of the r-th cardiac structure contour in the a-th cardiac section image is lower. When _{har1 -har2} is larger, it often indicates that the brightness of the r-th cardiac structure contour in the a-th cardiac section image is relatively higher than the brightness of its surrounding area, which often indicates that the cardiac structure represented by the r-th cardiac structure contour can be more clearly segmented, and often indicates that the recognizability of the r-th cardiac structure contour in the a-th cardiac section image is higher. Therefore, when E _a is larger, it often indicates that the recognizability of the cardiac structure contour in the a-th cardiac section image is relatively lower, which often indicates that the resolution of the a-th cardiac section image may be lower.

Step S3, determining a low-resolution factor corresponding to each heartbeat cycle according to the grayscale difference between each heartbeat cycle and the target echocardiogram in other heartbeat cycles, and the corresponding physical sign information of the patient to be tested.

As an example, the process of the method for determining the low resolution factor may be shown in FIG5 , and may specifically include the following steps:

Step 501, determining the information change amount between every two heartbeat cycles according to the grayscale difference between the target echocardiograms within every two heartbeat cycles.

For example, determining the amount of information change between every two heartbeat cycles may include the following steps:

In the first step, any heartbeat cycle is determined as a marking cycle, any heartbeat cycle other than the marking cycle is determined as a reference cycle, any frame of target echocardiogram within the marking cycle is determined as a marking image, and a target echocardiogram whose corresponding frame number is the same as the frame number corresponding to the marking image is screened out from the reference cycle as a reference image.

The serial number of the target echocardiogram collected in the heartbeat cycle to which it belongs is the frame serial number corresponding to the target echocardiogram. For example, the frame serial number corresponding to the first frame of the target echocardiogram collected in the heartbeat cycle may be 1. For another example, the frame serial number corresponding to the first frame of the target echocardiogram in the reference cycle may be equal to the frame serial number corresponding to the first frame of the target echocardiogram in the marking cycle.

In the second step, the absolute value of the difference between the average grayscale corresponding to the marked image and the average grayscale corresponding to the reference image is determined as the target grayscale difference between the marked image and the reference image.

The average grayscale corresponding to the marked image may be the average of the grayscale values corresponding to all pixels in the marked image. The average grayscale corresponding to the reference image may be the average of the grayscale values corresponding to all pixels in the reference image.

In the third step, the absolute value of the difference between the grayscale variance corresponding to the marked image and the grayscale variance corresponding to the reference image is determined as the grayscale distribution difference between the marked image and the reference image.

The grayscale variance corresponding to the marked image may be the variance of the grayscale values corresponding to all pixels in the marked image, and the grayscale variance corresponding to the reference image may be the variance of the grayscale values corresponding to all pixels in the reference image.

In the fourth step, a target difference factor between the marked image and the reference image is determined according to the target grayscale difference between the marked image and the reference image, and the grayscale distribution difference between the marked image and the reference image.

Among them, both the target grayscale difference and the grayscale distribution difference can be positively correlated with the target difference factor.

The fifth step is to determine the information change between every two heartbeat cycles based on all target difference factors within every two heartbeat cycles.

Among them, the target difference factor can be positively correlated with the amount of information change.

For example, the formula for determining the information change between any two heartbeat cycles can be:

Wherein, JB _fk is the information change between the f-th heart cycle and the k-th heart cycle. f and k are the serial numbers of different heart cycles. M is the total number of target echocardiograms within the heart cycle. m is the serial number of the target echocardiogram within the heart cycle. || is the absolute value function. _{h f} m is the average grayscale corresponding to the m-th frame target echocardiogram within the f-th heart cycle. _{h km} is the average grayscale corresponding to the m-th frame target echocardiogram within the k-th heart cycle. _{σ fm} is the grayscale variance corresponding to the m-th frame target echocardiogram within the f-th heart cycle. _{σ km} is the grayscale variance corresponding to the m-th frame target echocardiogram within the k-th heart cycle. |h _fm -h _km | is the target grayscale difference between the m-th frame target echocardiogram within the f-th heart cycle and the m-th frame target echocardiogram within the k-th heart cycle. |σ _fm -σ _km | is the grayscale distribution difference between the target echocardiogram of the mth frame in the fth heartbeat cycle and the target echocardiogram of the mth frame in the kth heartbeat cycle. |h _fm -h _km |×|σ _fm -σ _km | is the target difference factor between the target echocardiogram of the mth frame in the fth heartbeat cycle and the target echocardiogram of the mth frame in the kth heartbeat cycle.

It should be noted that when |h _fm -h _km | is larger, it often means that the overall grayscale presented between the target echocardiogram of the mth frame in the fth heartbeat cycle and the kth heartbeat cycle is less similar. When |σ _fm -σ _km | is larger, it often means that the grayscale distribution between the target echocardiogram of the mth frame in the fth heartbeat cycle and the kth heartbeat cycle is relatively less similar. Generally, different heartbeat cycles corresponding to the same patient in a short period of time are often similar, and the actual information distribution in different heartbeat cycles collected is often similar. When the grayscale of the ultrasound cardiograms taken in different heartbeat cycles varies greatly, it may often be due to the influence of factors such as the environment, resulting in different resolutions of the ultrasound cardiograms collected in different heartbeat cycles, thereby resulting in different grayscale distributions presented in the ultrasound cardiograms in different heartbeat cycles. When JB _fk is larger, it often means that the grayscale distribution changes between the f-th heartbeat cycle and the k-th heartbeat cycle are more dissimilar, it often means that the information change between the f-th heartbeat cycle and the k-th heartbeat cycle is larger, and it often means that there is more likely to be low-resolution echocardiograms in the f-th heartbeat cycle and the k-th heartbeat cycle.

Step 502: Determine a resolution evaluation value corresponding to the patient to be detected based on the physical sign information corresponding to the patient to be detected.

The physical sign information may be information related to the human body. For example, the physical sign information may include, but is not limited to, height, weight, heart position and size, chest wall structure, lung condition, blood flow condition, skin and soft tissue condition. The resolution evaluation value may indicate the possibility that the acquired patient echocardiogram has a low resolution.

It should be noted that the patient's physical signs often affect the resolution of the corresponding echocardiogram. For example, if the patient is obese, the resolution of the echocardiogram collected from the patient is often relatively low.

For example, the vital sign information corresponding to the patient to be detected may be input into a pre-trained resolution evaluation network, and the resolution evaluation value corresponding to the patient to be detected may be output through the pre-trained resolution evaluation network.

The resolution evaluation network may be a network used for resolution evaluation, for example, a FNN (Feedforward Neural Network, fully connected neural network).

For example, the training process of the resolution assessment network may include the following steps:

The first step is to build a resolution evaluation network.

For example, an FNN can be constructed as a resolution evaluation network before training.

The second step is to obtain a set of sample vital signs information.

Among them, the resolution evaluation value corresponding to the sample vital sign information in the above sample vital sign information set is known.

For example, multiple doctors can evaluate the possibility of low image resolution based on sample vital sign information, and determine the average of all evaluation values corresponding to the sample vital sign information as the resolution evaluation value corresponding to the sample vital sign information. The evaluation value obtained by the doctors can be in the range of [0, 1], and the larger the evaluation value, the greater the possibility of low resolution when acquiring the patient's echocardiogram.

In the third step, the constructed resolution assessment network is trained using the above sample vital sign information set as a training set and the resolution assessment value corresponding to the sample vital sign information as a training label to obtain a trained resolution assessment network.

Step 503: Determine a low resolution factor corresponding to each heartbeat cycle according to the resolution evaluation value and the information change amount between each heartbeat cycle and other heartbeat cycles.

Among them, both the information change amount and the resolution evaluation value can be positively correlated with the low resolution factor.

For example, the formula for determining the low-resolution factor corresponding to the heartbeat cycle may be:

Wherein, JG _f is the low resolution factor corresponding to the f-th heartbeat cycle. f and k are the serial numbers of different heartbeat cycles. N is the preset number, that is, the number of heartbeat cycles obtained. JB _fk is the information change between the f-th heartbeat cycle and the k-th heartbeat cycle. P is the resolution evaluation value corresponding to the patient to be tested.

It should be noted that, in general, when collecting the patient's echocardiogram, the doctor often places the ultrasound probe in a relatively suitable position. Therefore, the ultrasound cardiograms of the same patient with the same frame number in different heartbeat cycles are often similar. If the ultrasound cardiogram in a certain heartbeat cycle is less similar to the ultrasound cardiograms with the same frame number in other heartbeat cycles, it often indicates that the ultrasound cardiogram in the heartbeat cycle may be abnormal due to the influence of environmental factors. In actual situations, when the ultrasound cardiogram is abnormal, its resolution often becomes lower, and it often does not increase its resolution. And when P is larger, it often indicates that the resolution of the collected ultrasound cardiogram of the patient to be tested may be lower, and it often indicates that when upsampling the target ultrasound cardiogram, it is necessary to set a higher standard deviation for the Gaussian interpolation kernel. Therefore, when JG _f is larger, it often means that the echocardiogram in the f-th heart cycle is less similar to the echocardiograms in other heart cycles, and the resolution of the acquired echocardiogram of the patient to be tested may be lower; it often means that when upsampling the target echocardiogram in the f-th heart cycle, it is necessary to set a higher standard deviation for the Gaussian interpolation kernel.

Step S4, determining the target standard deviation corresponding to each frame of the target echocardiogram according to the identification clarity factor to be enhanced corresponding to the cardiac section image contained in each frame of the target echocardiogram and the low resolution factor corresponding to the heart cycle to which each frame of the target echocardiogram belongs.

As an example, the Euclidean norm between the clear identification factor to be enhanced corresponding to the cardiac section image contained in each frame of the target echocardiogram and the low resolution factor corresponding to the heart cycle to which each frame of the target echocardiogram belongs can be determined as the target standard deviation corresponding to each frame of the target echocardiogram.

Among them, the Euclidean norm is also called the Euclidean norm.

For example, the formula for determining the target standard deviation corresponding to the target echocardiogram may be:

Wherein, T _m is the target standard deviation corresponding to the target echocardiogram of the mth frame in the cardiac cycle. m is the serial number of the target echocardiogram in the cardiac cycle. E _m is the identification clear enhancement factor corresponding to the cardiac section image contained in the target echocardiogram of the mth frame in the cardiac cycle. JG _m is the low resolution factor corresponding to the cardiac cycle to which the target echocardiogram of the mth frame in the cardiac cycle belongs.

It should be noted that when _Em is larger, it often means that the recognizability of the cardiac structure contour in the cardiac section image contained in the target echocardiogram of the mth frame is relatively low, it often means that the resolution of the target echocardiogram of the mth frame may be lower, and it often means that a higher standard deviation needs to be set for the Gaussian interpolation kernel when upsampling the target echocardiogram of the mth frame. When _JGm is larger, it often means that the echocardiogram within the heartbeat cycle to which the target echocardiogram of the mth frame belongs is less similar to the echocardiograms within other heartbeat cycles, and the resolution of the acquired patient's ultrasound cardiogram may be lower; it often means that a higher standard deviation needs to be set for the Gaussian interpolation kernel when upsampling the target echocardiogram within the heartbeat cycle to which the target echocardiogram of the mth frame belongs, and it often means that a higher standard deviation needs to be set for the Gaussian interpolation kernel when upsampling the target echocardiogram of the mth frame. Therefore, _Tm can represent the standard deviation that needs to be set for the Gaussian interpolation kernel when upsampling the target echocardiogram of the mth frame.

Step S5, upsampling each frame of the target echocardiogram through a Gaussian interpolation kernel according to the target standard deviation corresponding to each frame of the target echocardiogram to obtain a target image.

The target standard deviation is the standard deviation obeyed by the Gaussian interpolation kernel. Upsampling is also called upsampling interpolation or image pyramid algorithm.

As an example, when upsampling any frame of target echocardiogram, the target standard deviation corresponding to the target echocardiogram can be used as the standard deviation that the Gaussian interpolation kernel needs to obey. Through the Gaussian interpolation kernel, upsampling of the target echocardiogram can be achieved, and the image obtained after upsampling can be determined as the target image. The doctor can make a preliminary diagnosis of the cardiovascular condition of the patient to be tested by observing all the target images.

Specifically, according to the target standard deviation corresponding to each frame of the target echocardiogram, upsampling each frame of the target echocardiogram through the Gaussian interpolation kernel may include: for all frames of the target echocardiogram within a heartbeat cycle, blank cells may be added to adjacent rows and columns in the image matrix, and the size of the image is ensured to be consistent, and the Gaussian kernel is used as the interpolation kernel in the upsampling interpolation process, and T _m is used as the standard deviation obeyed by the interpolation kernel of the mth frame of the target echocardiogram within the heartbeat cycle. The Gaussian kernel will generate a set of convolution kernel weights that obey the standard deviation, and then use them as weight values for convolving the blank cells inserted into the original image with the neighboring pixel points, thereby obtaining the pixel values of the pixel points represented by the inserted blank cells, thereby completing the upsampling interpolation of each frame of the target echocardiogram to enhance the resolution of the target echocardiogram.

Referring to FIG6 , based on the same inventive concept as the above method embodiment, the present invention provides a cardiovascular condition auxiliary detection system for internal medicine, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the above computer program is executed by the processor, the steps of implementing a cardiovascular condition auxiliary detection system for internal medicine may specifically include:

An image acquisition module 601 is used to acquire a target echocardiogram of a patient to be tested at each preset time within a preset number of heartbeat cycles, wherein the target echocardiogram includes a heart section image;

A screening and determination module 602 is used to screen out the cardiac structure contour from each cardiac section image, and determine the identification clarity enhancement factor corresponding to each cardiac section image according to the number of edge lines on the cardiac structure contour in each cardiac section image and the grayscale difference between the pixels on the cardiac structure contour;

The resolution factor determination module 603 is used to determine the low resolution factor corresponding to each heartbeat cycle according to the grayscale difference between each heartbeat cycle and the target echocardiogram in other heartbeat cycles, and the physical sign information corresponding to the patient to be detected;

The standard deviation determination module 604 is used to determine the target standard deviation corresponding to each frame of the target echocardiogram according to the identification clarity factor to be enhanced corresponding to the cardiac section image contained in each frame of the target echocardiogram and the low resolution factor corresponding to the heart cycle to which each frame of the target echocardiogram belongs;

The up-sampling module 605 is used to up-sample each frame of the target echocardiogram through a Gaussian interpolation kernel according to a target standard deviation corresponding to each frame of the target echocardiogram to obtain a target image, wherein the target standard deviation is the standard deviation obeyed by the Gaussian interpolation kernel.

Fig. 7 is a schematic diagram of the structure of a computer device provided by an embodiment of the present invention. Exemplarily, as shown in Fig. 7, the computer device 700 includes: a memory 701, a processor 702, and a computer program 703 stored in the memory 701 and running on the processor 702, wherein when the processor 702 executes the computer program 703, the computer device can execute any of the cardiovascular condition auxiliary detection methods for internal medicine described above.

Based on the same inventive concept as the above method embodiment, the present invention provides a server, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes any one of the above cardiovascular condition auxiliary detection methods for internal medicine.

Based on the same inventive concept as the above method embodiment, the present invention provides a computer program product, which includes: computer program code, when the computer program code runs on a computer, the computer executes any one of the above cardiovascular condition auxiliary detection methods for internal medicine.

Based on the same inventive concept as the above-mentioned method embodiment, the present invention provides a computer-readable storage medium, which stores a computer program code. When the computer program code runs on a computer, the computer executes any one of the above-mentioned cardiovascular condition auxiliary detection methods for internal medicine.

In summary, the present invention comprehensively considers multiple indicators related to resolution, such as identifying clear factors to be enhanced and low resolution factors, and quantifies the target standard deviation required for upsampling each frame of target echocardiogram, so that adaptive upsampling of each frame of target echocardiogram can be achieved, thereby improving the upsampling effect. Secondly, the process of quantifying the target standard deviation of the present invention is relatively objective, which reduces the influence of human subjective factors to a certain extent, thereby improving the accuracy of determining the target standard deviation, and then improving the effect of upsampling each frame of target echocardiogram.

The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. 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 can still be modified, or some of the technical features can be replaced by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the 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.

Claims (10)

1. A method for assisted detection of cardiovascular conditions in medical medicine, comprising the steps of:

Acquiring a target echocardiogram of a patient to be detected at each preset time in a preset number of heartbeat cycles, wherein the target echocardiogram comprises a heart section image;

Screening out the outline of the heart structure from each heart section image, and determining the identification clear factor to be enhanced corresponding to each heart section image according to the number of edge lines on the outline of the heart structure in each heart section image and the gray level difference between pixel points on the outline of the heart structure;

determining a low resolution factor corresponding to each heartbeat cycle according to gray level difference between each heartbeat cycle and the target echocardiogram in other heartbeat cycles and sign information corresponding to the patient to be detected;

Determining a target standard deviation corresponding to each frame of target echocardiogram according to the identification clear to-be-enhanced factor corresponding to the heart section image contained in each frame of target echocardiogram and the low resolution factor corresponding to the heartbeat period to which each frame of target echocardiogram belongs;

And up-sampling each frame of the target echocardiogram through a Gaussian interpolation kernel according to the target standard deviation corresponding to each frame of the target echocardiogram to obtain a target image, wherein the target standard deviation is the standard deviation obeyed by the Gaussian interpolation kernel.

2. The method for assisting in detecting a cardiovascular condition for internal medicine according to claim 1, wherein the determining the identifying clear to-be-enhanced factor corresponding to each heart section image according to the number of edge lines on the heart structure outline in each heart section image and the gray level difference between pixel points on the heart structure outline comprises:

determining the absolute value of the difference value of gray values corresponding to every two pixel points on each heart structure outline as the reference gray difference between the two pixel points;

determining gray scale representation differences corresponding to each heart structure outline according to reference gray scale differences among all pixel points on each heart structure outline, wherein the reference gray scale differences and the gray scale representation differences are in positive correlation;

Performing edge detection on each heart structure outline to obtain edge lines, and determining an edge identification index to be adjusted corresponding to each heart structure outline according to the number of the edge lines on each heart structure outline and the corresponding gray scale representation differences, wherein the number of the edge lines on the heart structure outline and the corresponding gray scale representation differences are in positive correlation with the edge identification index to be adjusted corresponding to the heart structure outline;

And determining the identification clear factor to be enhanced corresponding to each heart section image according to the edge identification index to be adjusted corresponding to all heart structure outlines in each heart section image and the gray value corresponding to the pixel point on each heart structure outline in each heart section image.

3. The method for assisting in detecting a cardiovascular condition in a medical department according to claim 2, wherein the determining the identifying clear to-be-enhanced factor corresponding to each heart section image according to the edge identifying to-be-adjusted index corresponding to all heart structure outlines in each heart section image and the gray value corresponding to the pixel point on each heart structure outline in each heart section image comprises:

Determining a union set of preset neighborhoods corresponding to all pixel points on each heart structure outline as a reference neighborhood corresponding to each heart structure outline;

Identifying the index to be adjusted according to each heart structure contour and the corresponding reference neighborhood and edge thereof, and determining the contour index to be adjusted corresponding to each heart structure contour;

And determining the identification clear to-be-enhanced factors corresponding to each heart section image according to the outline to-be-adjusted indexes corresponding to all heart structure outlines in each heart section image, wherein the outline to-be-adjusted indexes and the identification clear to-be-enhanced factors are in positive correlation.

4. A method for assisting in detecting a cardiovascular condition for medical use according to claim 3, wherein the determining the outline to be adjusted index corresponding to each heart structure outline according to each heart structure outline and the corresponding reference neighborhood and edge identification to be adjusted index comprises:

Determining an average value of gray values corresponding to all pixel points on each heart structure outline as a first gray representing factor corresponding to each heart structure outline;

determining the average value of gray values corresponding to all pixel points in a reference neighborhood corresponding to each heart structure outline as a second gray representing factor corresponding to each heart structure outline;

Determining the difference between the first gray scale representative factor and the second gray scale representative factor corresponding to each heart structure outline as the gray scale contrast difference corresponding to each heart structure outline;

And determining the outline to-be-adjusted index corresponding to each heart structure outline according to the edge identification to-be-adjusted index corresponding to each heart structure outline and the gray scale comparison difference, wherein the edge identification to-be-adjusted index and the outline to-be-adjusted index form a positive correlation, and the gray scale comparison difference and the outline to-be-adjusted index form a negative correlation.

5. The method for assisting in detecting cardiovascular conditions in internal medicine according to claim 1, wherein determining the low resolution factor corresponding to each heartbeat cycle according to the gray scale difference between the target echocardiogram in each heartbeat cycle and other heartbeat cycles and the sign information corresponding to the patient to be detected comprises:

According to the gray level difference between the target echocardiogram in every two heartbeat periods, determining the information change quantity between every two heartbeat periods;

determining a resolution evaluation value corresponding to the patient to be detected according to the sign information corresponding to the patient to be detected;

and determining a low resolution factor corresponding to each heartbeat cycle according to the resolution evaluation value and the information change quantity between each heartbeat cycle and other heartbeat cycles, wherein the information change quantity and the resolution evaluation value are in positive correlation with the low resolution factor.

6. The method for assisted detection of cardiovascular conditions for medical use according to claim 5, wherein the determining the amount of information change between each two heart beat cycles based on the gray level difference between the target echocardiogram for each two heart beat cycles comprises:

Determining any one heartbeat period as a marking period, determining any one heartbeat period outside the marking period as a reference period, determining any one frame of target echocardiogram in the marking period as a marking image, and screening out the target echocardiogram with the same corresponding frame number as the frame number corresponding to the marking image from the reference period as a reference image;

determining the absolute value of the difference between the average gray scale corresponding to the mark image and the average gray scale corresponding to the reference image as the target gray scale difference between the mark image and the reference image;

determining the absolute value of the difference between the gray variance corresponding to the mark image and the gray variance corresponding to the reference image as the gray distribution difference between the mark image and the reference image;

Determining a target difference factor between the mark image and the reference image according to a target gray level difference between the mark image and the reference image and a gray level distribution difference between the mark image and the reference image, wherein the target gray level difference and the gray level distribution difference are in positive correlation with the target difference factor;

And determining the information change quantity between every two heartbeat cycles according to all target difference factors in every two heartbeat cycles, wherein the target difference factors and the information change quantity form a positive correlation.

7. The method for assisted detection of cardiovascular conditions for medical use according to claim 5, wherein the determining the resolution evaluation value corresponding to the patient to be detected according to the sign information corresponding to the patient to be detected comprises:

inputting the physical sign information corresponding to the patient to be detected into a pre-trained resolution evaluation network, and outputting a resolution evaluation value corresponding to the patient to be detected through the pre-trained resolution evaluation network, wherein the training process of the resolution evaluation network comprises the following steps:

constructing a resolution evaluation network;

acquiring a sample sign information set, wherein a resolution evaluation value corresponding to sample sign information in the sample sign information set is known;

And training the constructed resolution evaluation network by taking the sample sign information set as a training set and the resolution evaluation value corresponding to the sample sign information as a training label to obtain a trained resolution evaluation network.

8. The method for assisting in detecting a cardiovascular condition for internal medicine according to claim 1, wherein the determining the target standard deviation corresponding to each frame of target echocardiogram according to the recognition clear to-be-enhanced factor corresponding to the heart section image included in each frame of target echocardiogram and the low resolution factor corresponding to the heartbeat period to which each frame of target echocardiogram belongs comprises:

And determining the Euro norm between the recognized clear factors to be enhanced corresponding to the heart section images contained in each frame of target echocardiogram and the low resolution factors corresponding to the heartbeat period to which each frame of target echocardiogram belongs as the target standard deviation corresponding to each frame of target echocardiogram.

9. The method for assisted detection of cardiovascular conditions for medical use according to claim 1, wherein said screening out heart structure contours from each of the heart cut images comprises:

Inputting each heart section image into a pre-trained heart structure contour recognition network, and recognizing the heart structure contour in each heart section image through the pre-trained heart structure contour recognition network.

10. The method for assisted detection of cardiovascular conditions for medical use of claim 9, wherein the training process of the cardiac structure profile recognition network comprises:

Constructing a heart structure contour recognition network;

Acquiring a sample heart section image set, wherein heart structure outlines are marked in sample heart section images in the sample heart section image set;

And training the constructed heart structure contour recognition network by taking the sample heart section image set as a training set and taking the heart structure contour marked in the sample heart section image as a training label to obtain a trained heart structure contour recognition network.