A method, system, and medium for automatic measurement of a cardiac doppler ultrasound spectrogram image

Abstract

Some embodiments of the specification provide an automatic measurement method of a heart Doppler ultrasound spectrum image, comprising: acquiring a continuously updated current frame ultrasound spectrum image according to a preset time interval, and processing the current frame ultrasound spectrum image to obtain one or more to-be-processed subgraphs of blood flow peaks; inputting each to-be-processed subgraph into an evaluation and mask determination model to obtain a quality score and a mask image of each to-be-processed subgraph; based on the quality score of the one or more to-be-processed subgraphs, screening one or more target subgraphs from the one or more to-be-processed subgraphs; based on the mask image of the one or more target subgraphs, performing contour extraction on each target subgraph to obtain a contour of the blood flow peak in each target subgraph; and based on the contour of the blood flow peak in the one or more target subgraphs, calculating a target parameter and showing the target parameter to a user.

Description

Technical Field

[0001] This application generally relates to the field of medical technology, and in particular to an automatic measurement method for cardiac Doppler ultrasound spectral images. Background Technology

[0002] Cardiac hemodynamic parameters, such as peak velocity (Vmax), velocity-time integral (VTI), stroke volume (SV), and cardiac output (CO), are important indicators for clinical monitoring of cardiac function. Stroke volume reflects the pumping efficiency of a single heartbeat; a decrease suggests heart failure or shock, while an increase may be due to compensation or pharmacological effects. Cardiac output assesses overall cardiac pumping efficiency and is directly related to tissue oxygenation; abnormalities can lead to insufficient organ perfusion. Monitoring cardiac hemodynamic parameters is crucial for shock classification, treatment evaluation, and critical care management, guiding precise treatment (such as fluid resuscitation or adjustment of cardiotonics), especially in scenarios like heart failure and major surgery, enabling early detection of circulatory failure and prevention of multi-organ damage. In clinical practice, pulsed Doppler ultrasound is commonly used to measure cardiac hemodynamic parameters due to its simplicity, efficiency, and non-invasiveness. In a typical Doppler ultrasound measurement, doctors usually need to manually drag the time axis to find the optimal blood flow peak and outline its contour in order to calculate parameters such as Vmax and VTI.

[0003] In actual examinations, this method, which relies entirely on manual measurement by doctors, is quite complex, requires a high level of expertise from the doctors, and is very time-consuming. Furthermore, a single ultrasound spectral image acquisition yields a series of blood flow peaks; selecting only one is not representative and makes it impossible to calculate parameters such as heart rate. If doctors manually select multiple peaks for measurement, the time required will increase exponentially.

[0004] Therefore, an automatic measurement method for cardiac Doppler ultrasound spectral images is proposed, which can automatically identify the position of blood flow peaks, automatically detect peak values ​​and contours, and improve scanning efficiency. Summary of the Invention

[0005] This specification provides an automatic measurement method for cardiac Doppler ultrasound spectral images, comprising: acquiring continuously updated current frame ultrasound spectral images at preset time intervals, processing the current frame ultrasound spectral images to obtain one or more sub-images to be processed, representing blood flow peaks; inputting each sub-image to be processed into an evaluation and mask determination model to obtain a quality score and mask image for each sub-image; selecting one or more target sub-images from the one or more sub-images to be processed based on the quality score; extracting the contour of each target sub-image based on the mask image of the one or more target sub-images to obtain the contour of the blood flow peaks in each target sub-image; and calculating target parameters based on the contours of the blood flow peaks in the one or more target sub-images and displaying the target parameters to the user.

[0006] In some embodiments, the method further includes: in response to the user performing a freeze operation, determining a target ultrasound spectrum image and displaying it to the user based on the start timestamps of all blood flow peaks in all acquired target sub-images and their corresponding quality scores, wherein the user confirms whether a freeze operation needs to be performed based on the target parameters of all blood flow peaks in all acquired target sub-images.

[0007] In some embodiments, processing the current frame ultrasound spectrum image to obtain one or more sub-images of blood flow peaks includes: segmenting the current frame ultrasound spectrum image based on the horizontal axis to obtain an intermediate ultrasound spectrum image below the horizontal axis; binarizing and extracting the envelope of the intermediate ultrasound spectrum image to determine the intermediate curve and smoothing it; cropping the intermediate ultrasound spectrum image based on multiple minimum value positions in the intermediate curve to obtain single-peak sub-images of each single peak, and calculating the start timestamp of each single-peak sub-image; and determining whether a single-peak sub-image is a sub-image to be processed based on the start timestamp of each single-peak sub-image.

[0008] In some embodiments, determining whether a single-peak sub-image is a sub-image to be processed based on the start timestamp of each single-peak sub-image includes: determining single-peak sub-images with start timestamps greater than reference timestamps as sub-images to be processed, wherein the reference timestamp is the peak timestamp of the blood flow peak in the last target sub-image corresponding to the previous frame of ultrasound spectrum image.

[0009] In some embodiments, selecting one or more target subgraphs from one or more subgraphs to be processed based on the quality scores of one or more subgraphs to be processed includes: identifying subgraphs to be processed with quality scores greater than or equal to a score threshold as target subgraphs.

[0010] In some embodiments, in response to a user performing a freeze operation, a target ultrasound spectrum image is determined and displayed to the user based on the start timestamps of all blood flow peaks in all acquired target sub-images and their corresponding quality scores, including: stopping the updating of the current frame ultrasound spectrum image in response to the user performing a freeze operation; sorting all blood flow peaks in all acquired target sub-images in chronological order, and performing the following operations on all blood flow peaks in sequence: dividing all blood flow peaks into frames using a time window, the window length of which is the time length corresponding to the width of a single frame ultrasound spectrum image; determining the target blood flow peak within each frame signal, the target blood flow peak being the blood flow peak completely included within the frame signal; calculating the comprehensive quality score of the target blood flow peak, the comprehensive quality score being the sum of the quality scores of all target sub-images corresponding to the target blood flow peaks within the frame signal; and determining the frame signal with the largest comprehensive quality score as the target ultrasound spectrum image.

[0011] In some embodiments, after determining the target ultrasound spectrum image, the method further includes: updating the target parameters based on the target parameters corresponding to the target blood flow peaks within the target ultrasound spectrum image.

[0012] In some embodiments, the method further includes updating a preset time interval based on the calculated cardiac cycle of the subject.

[0013] In some embodiments, the target parameters include at least one of peak velocity Vmax, time-velocity integral VTI, stroke volume SV, peak start time stamp, peak end time stamp, heart rate HR, cardiac output CO, quality score, and peak profile.

[0014] This specification also provides an automatic measurement system for cardiac Doppler ultrasound spectral images, including a processor for executing the above-described automatic measurement method for cardiac Doppler ultrasound spectral images.

[0015] This specification also provides a computer-readable storage medium. This storage medium stores computer instructions, and when a computer reads the computer instructions from the storage medium, the computer executes the above-described automatic measurement method for cardiac Doppler ultrasound spectral images. Attached Figure Description

[0016] This application will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:

[0017] Figure 1 These are schematic diagrams illustrating application scenarios of exemplary ultrasound imaging systems according to some embodiments of this specification; Figure 2These are exemplary block diagrams of an ultrasound imaging system according to some embodiments of this specification; Figure 3 This is an exemplary flowchart illustrating an automatic measurement method for cardiac Doppler ultrasound spectral images according to some embodiments of this specification; Figure 4 This is an exemplary flowchart illustrating a method for determining a subgraph to be processed according to some embodiments of this specification; Figure 5 This is a schematic diagram illustrating the determination of the quality score and mask image for each sub-image to be processed using an exemplary evaluation and mask determination model according to some embodiments of this specification; Figure 6 This is a schematic diagram illustrating the output results of a single calculation according to some embodiments of this specification; Figure 7A This is an exemplary flowchart illustrating the determination of a subgraph to be processed according to some embodiments of this specification; Figure 7B This is a schematic diagram illustrating the quality scoring of the subgraph to be processed according to some embodiments of this specification; and Figure 7C This is a schematic diagram showing the results of sorting all blood flow peaks by time according to some embodiments of this specification. Detailed Implementation

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0019] It should be understood that the terms “system,” “device,” “unit,” and / or “module” used herein are one way of distinguishing different levels of components, elements, parts, sections, or assemblies. However, if other words can achieve the same purpose, they may be replaced by other expressions.

[0020] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0021] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0022] The wearable devices described in the embodiments of this specification will be described in detail below with reference to the accompanying drawings. It should be noted that the following embodiments are only used to explain this specification and do not constitute a limitation thereof.

[0023] Figure 1 These are schematic diagrams illustrating application scenarios of exemplary ultrasound imaging systems according to some embodiments of this specification. For example... Figure 1 As shown, the ultrasound imaging system 100 may include a medical device 110, a processing device 120, a storage device 130, a terminal device 140, and a network 150. The various components of the ultrasound imaging system 100 can be interconnected and / or communicate with each other via wireless connection, wired connection, or a combination of both.

[0024] The medical device 110 can be configured to acquire ultrasound imaging data associated with at least one site of a subject to generate an image of that at least one site. In some embodiments, the ultrasound imaging data can be two-dimensional (2D) imaging data, three-dimensional (3D) imaging data, four-dimensional (4D) imaging data, etc., or any combination thereof. The subject to be examined can be a biological object or a non-biological object. For example, the subject to be examined may include a patient, an artificial object, etc. As another example, the subject to be examined may include a specific site, organ, and / or tissue of a patient. For example, the subject to be examined may include the heart, thyroid gland, esophagus, trachea, stomach, gallbladder, small intestine, colon, bladder, uterus, fallopian tubes, etc., or any combination thereof.

[0025] In some embodiments, medical device 110 may be a medical imaging device based on ultrasound patterns, such as a Doppler ultrasound diagnostic instrument, an ultrasound diagnostic instrument, or an ultrasound Doppler blood flow analyzer. The medical device 110 described above is merely illustrative and does not constitute a limitation on the scope of protection. Medical device 110 can acquire medical ultrasound images related to blood flow velocity based on the Doppler effect. In some embodiments, medical device 110 can acquire a patient's medical ultrasound image (e.g., a pulsed Doppler ultrasound spectral image) and send the medical ultrasound image to processing device 120. In some embodiments, medical device 110 can interact with other components of ultrasound imaging system 100 (e.g., processing device 120, storage device 130, terminal device 140) via network 150 for data and / or information exchange. In some embodiments, one or more components of ultrasound imaging system 100 (e.g., processing device 120, storage device 130, terminal device 140) may be integrated into medical device 110.

[0026] Processing device 120 can process information and / or data obtained from medical device 110, terminal device 140, and / or storage device 130. For example, processing device 120 can acquire continuously updated current frame ultrasound spectrum images at preset time intervals and process the current frame ultrasound spectrum images to obtain sub-images of blood flow peaks to be processed. Further, processing device 120 can use an evaluation and segmentation model to filter out target sub-images from the sub-images to be processed and determine the contours of blood flow peaks in each target sub-image. Processing device 120 can calculate and display target parameters to the user based on the contours of blood flow peaks in the target sub-images. In some embodiments, processing device 120 can be a computer, a user control platform, a single server, or a group of servers, etc. The server group can be centralized or distributed. In some embodiments, processing device 120 can be local or remote. In some embodiments, processing device 120 can be implemented on a cloud platform. For example, the cloud platform can include one or more of the following: private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, cross-cloud, multi-cloud, etc. In some embodiments, the processing device 120 may be part of the medical device 110 or the terminal device 140.

[0027] Storage device 130 may be configured to store data and / or instructions. This data and / or instructions may be obtained from processing device 120, medical device 110, and / or any other component of ultrasound imaging system 100. In some embodiments, storage device 130 may store data and / or instructions executable by processing device 120 or used to implement the exemplary methods described herein. In some embodiments, storage device 130 may include mass storage devices, removable storage devices, volatile read-write memory, read-only memory (ROM), etc., or any combination thereof. In some embodiments, storage device 130 may be implemented on a cloud platform as described in other parts of this specification.

[0028] Terminal device 140 may be configured to receive information and / or data from processing device 120, medical device 110, and / or storage device 130 via network 150. For example, terminal device 140 may receive images from processing device 120. In some embodiments, terminal device 140 may provide a user interface through which a user can view information and / or input data and / or instructions to ultrasound imaging system 100. For example, a user can view information related to medical device 110 through the user interface. As another example, a user can input user-input instructions through the user interface to adjust the update frequency of the current frame ultrasound spectral image. In some embodiments, terminal device 140 may include mobile device 140. 1. Tablet computer 140 2. Laptop computer 140 3, or any combination thereof. In some embodiments, the terminal device 140 may include a display capable of displaying information in a human-readable form, such as text, images, audio, video, charts, animations, etc., or any combination thereof.

[0029] Network 150 facilitates information and / or data exchange within the ultrasound imaging system 100. In some embodiments, one or more components of the ultrasound imaging system 100 (e.g., medical device 110, processing device 120, terminal device 140, or storage device 130) may transmit information and / or data to one or more other components of the ultrasound imaging system 100 via network 150. In some embodiments, network 150 may be any type of wired or wireless network, or a combination thereof.

[0030] It should be noted that the above description of the ultrasound imaging system 100 is for illustrative purposes only and is not intended to limit the scope of this application. It is understood that those skilled in the art, after understanding the principles of the system, can make various modifications and changes in form and detail to the application areas implementing the above system without departing from these principles. However, these changes and modifications do not depart from the scope of this application. For example, the ultrasound imaging system 100 may include one or more additional components, and / or one or more of the aforementioned components may be omitted. As another example, two or more components of the ultrasound imaging system 100 may be integrated into a single component. Yet another example is that a component of the ultrasound imaging system 100 may be replaced by another component capable of performing the function of that component. However, such variations and modifications do not depart from the protection scope of this specification.

[0031] Figure 2 This is an exemplary block diagram of an ultrasound imaging system according to some embodiments of this specification. In some embodiments, the ultrasound imaging system 200 may be implemented by a processing device 120. Figure 2 As shown, the ultrasound imaging system 200 may include an acquisition module 210, a sub-image determination module 220, a target sub-image determination module 230, and a target parameter determination module 240.

[0032] The acquisition module 210 can be configured to acquire continuously updated current frame ultrasound spectrum images at preset time intervals.

[0033] The sub-image determination module 220 can be configured to process the current frame ultrasound spectrum image to obtain one or more sub-images of blood flow peaks.

[0034] The target sub-image determination module 230 can be configured to input each sub-image to be processed into the evaluation and mask determination model to obtain a quality score and mask image for each sub-image to be processed; and based on the quality scores of multiple sub-images to be processed, to select one or more target sub-images from the multiple sub-images to be processed.

[0035] The target parameter determination module 240 can be configured to extract contours from one or more target sub-images using mask images to obtain the contours of blood flow peaks in each target sub-image; and to calculate target parameters based on the contours of blood flow peaks in one or more target sub-images, and display the target parameters to the user. In some embodiments, the target parameter determination module 240 can also be configured to, in response to a user performing a freeze operation, determine a target ultrasound spectral image based on the start timestamps of all blood flow peaks in all acquired target sub-images and their corresponding quality scores, and display it to the user.

[0036] Each of the above modules may be a hardware circuit designed to perform a specific operation (e.g., according to a set of instructions stored in one or more storage media) and / or any combination of the hardware circuit and the one or more storage media.

[0037] It should be noted that the above description of the system and its modules is for convenience only and should not limit this application to the scope of the embodiments described. It is understood that those skilled in the art, after understanding the principle of the system, may arbitrarily combine the modules or construct subsystems connected to other modules without departing from this principle. For example, in some embodiments, the subgraph determination module 220 and the target subgraph determination module 230 can be integrated into one module. As another example, all modules can share a storage device, or each module can have its own independent storage device. Such variations are all within the protection scope of this application.

[0038] Figure 3 This is an exemplary flowchart illustrating an automated measurement method for cardiac Doppler ultrasound spectral images according to some embodiments of this specification. In some embodiments, process 300 may be executed automatically by ultrasound imaging system 100. For example, process 300 may be implemented as a set of instructions (e.g., an application program) stored in storage device 130, processing device 120, and / or Figure 2 The module in the document can execute this set of instructions and perform process 300 accordingly. The operation of the process shown below is intended to be illustrative. In some embodiments, process 300 may be completed by one or more additional operations not described and / or not by one or more operations discussed in this specification. Additionally, as... Figure 3 The order of operations shown and described below is not restrictive.

[0039] In step 310, the processing device 120 may acquire continuously updated current frame ultrasound spectrum images at preset time intervals. In some embodiments, step 310 may be performed by the acquisition module 210 in the system 200.

[0040] Ultrasound spectral images refer to images acquired through techniques such as echocardiography. In echocardiography, the Doppler effect is used to measure velocity information such as blood flow velocity and myocardial motion velocity within the heart, and this velocity information is displayed in the form of a spectral image (i.e., a Doppler ultrasound spectral image). The horizontal axis (also called the x-axis or time axis) of the ultrasound spectral image represents the sampling time, with each point on the horizontal axis representing a time point; the vertical axis (also called the y-axis or velocity axis) represents the motion velocities related to the heart (such as blood flow velocity or myocardial motion velocity), with each point on the vertical axis representing a velocity value.

[0041] In some embodiments, ultrasound spectrum images can be acquired from a medical device (e.g., medical device 110), storage device 130, or any other storage device at preset time intervals. For example, during an ultrasound examination, the ultrasound probe of the medical device (or ultrasound imaging system 100) continuously emits ultrasound waves toward the heart. When the ultrasound waves encounter moving blood flow or myocardial tissue within the heart, they are reflected. After receiving the reflected ultrasound waves, the medical device can send the reflected ultrasound waves to the processing device 120 in real time or periodically. The processing device 120 can calculate velocity information based on the Doppler frequency shift principle, continuously generate scrolling ultrasound spectrum images, and store them in the storage device 130 for retrieval at preset time intervals, thereby updating the current frame of the ultrasound spectrum image. In this case, in some embodiments, the preset time interval can be less than 500 ms (e.g., the preset time interval can be 400 ms, 300 ms, 200 ms, etc.). This setting ensures that the processing device 120 can calculate and display target parameters in real time based on the ultrasound data acquired from the user for the doctor's diagnosis. For example, storage device 130 or any other storage device may pre-store multiple ultrasound spectrum images ordered by time. Processing device 120 may sequentially retrieve ultrasound spectrum images from storage device 130 or any other storage device as the current frame ultrasound spectrum image according to a preset time interval.

[0042] In some embodiments, the processing device 120 may update a preset time interval. For example, the processing device 120 may update the preset time interval based on the cardiac cycle (or heart rate) of the subject calculated below. The preset time interval may reflect the frequency at which the processing device 120 updates / acquires the current frame spectral image. The preset time interval (the calculation frequency of the decision processing device 120) is updated in real time according to the heart rate of the subject; a faster heart rate shortens the preset time interval (i.e., increases the calculation frequency), while a slower heart rate lengthens the preset time interval (i.e., decreases the calculation frequency). This ensures that the processing device 120 updates information with each calculation and that newly appearing blood flow peaks are calculated promptly, while reducing unnecessary calculations, improving the efficiency of real-time computation, and reducing performance overhead.

[0043] For example, in the initial stage of an ultrasound examination, the medical equipment has not yet acquired enough ultrasound spectral image signals. At this time, the preset time interval can be set to be relatively short (e.g., 2 seconds) to ensure that there are two or more cardiac cycles in the current frame of the ultrasound spectral image, facilitating the first calculation of the cardiac cycle. It is important to note that the preset time interval should not be too short initially, because if there is only one cardiac cycle (i.e., only one blood flow peak) in the current frame of the ultrasound spectral image, the interval between adjacent blood flow peaks cannot be calculated, meaning the cardiac cycle and heart rate cannot be calculated. Subsequently, the processing device 120 can update the preset time interval based on the calculated cardiac cycle, ensuring that at least one new complete cardiac cycle appears in the next calculation.

[0044] As an example, in the initial stage of an ultrasound examination, when the preset time interval is set to 2 seconds (equivalent to updating the current frame every 2 seconds), and the length of each ultrasound spectrum image frame is 4 seconds, then the first frame is the ultrasound spectrum image from 0 to 2 seconds, with 2 to 4 seconds being blank; the second frame is the ultrasound spectrum image from 0 to 4 seconds; the third frame is the ultrasound spectrum image from 2 to 6 seconds (the entire window can only display a length of 4 seconds), and so on. When the ultrasound spectrum image does not include blank images, the interval time can be reset according to the obtained heart rate, for example, 0.8 seconds or 1 second, to ensure that a new complete cardiac cycle appears each time, balancing information update frequency and computational overhead. In other embodiments, to reduce unnecessary calculations, the preset time interval can also be appropriately extended (for example, set to 3 seconds, 3.5 seconds, etc.) to reduce the overlap between two adjacent ultrasound spectrum images.

[0045] In step 320, the processing device 120 can process the current frame of ultrasound spectral image to obtain one or more sub-images of blood flow peaks to be processed. In some embodiments, step 320 can be performed by the sub-image determination module 220 in system 200.

[0046] The sub-image to be processed refers to a local image containing each peak, cropped from the current frame of the ultrasound spectrum image according to the position of each peak. In some embodiments, the processing device 120 may determine the sub-image to be processed corresponding to the current frame of the ultrasound spectrum image according to the following operations. Figure 4 This is an exemplary flowchart illustrating a method for determining a subgraph to be processed according to some embodiments of this specification.

[0047] In step 410, the processing device 120 can segment the current frame ultrasound spectrum image based on the horizontal axis to obtain an intermediate ultrasound spectrum image below the horizontal axis. In other words, the intermediate ultrasound spectrum image can mainly contain velocity information related to blood flow peaks.

[0048] In some embodiments, the ultrasound spectral image in a complete cardiac examination record may contain multiple cardiac cycles and other information that may interfere with the analysis. The purpose of axis-based segmentation is to extract the intermediate ultrasound spectral image below the axis, thereby making subsequent analysis of blood flow peaks more accurate.

[0049] In step 420, the processing device 120 may binarize the intermediate ultrasound spectrum image to determine the intermediate curve. In this specification, the intermediate curve refers to a curve representing velocity changes (e.g., blood flow velocity) related to cardiac activity displayed in the ultrasound spectrum image.

[0050] In some embodiments, the processing device 120 may first binarize the intermediate ultrasound spectrum image. The processing device 120 may then perform an envelope extraction operation on the binarized image to obtain an envelope. Exemplary envelope extraction algorithms include Hilbert-Huang transform (HHT), wavelet transform envelope method, extreme point spline interpolation method, etc. Further, the processing device 120 may extract the ordinate of the envelope to obtain an intermediate curve. For example, the obtained envelope may encompass the entire blood flow region. For each coordinate on the horizontal axis, it may correspond to one or more points on the envelope. The processing device 120 may retain only the ordinate corresponding to the lower envelope points, thus obtaining the intermediate curve, which corresponds to the changing trend of the outermost layer of blood flow.

[0051] In some embodiments, the processing device 120 may employ a smoothing algorithm to smooth the envelope. Exemplary smoothing algorithms may include moving average algorithms, Gaussian filtering algorithms, median filtering algorithms, bilateral filtering algorithms, etc. For example, the processing device 120 may use a moving average algorithm to smooth the envelope. After smoothing, noise in the intermediate ultrasound spectrum image can be reduced, making the generated intermediate curve smoother and facilitating accurate identification of minimum value locations. This is because the ultrasound spectrum image may experience small fluctuations due to factors such as equipment signal interference, and smoothing can eliminate these interferences.

[0052] In step 430, the processing device 120 can crop the intermediate ultrasound spectrum image based on multiple minimum value positions in the intermediate curve to obtain single-peak sub-images of each single peak, and calculate the start timestamp of each single-peak sub-image. The minimum value position refers to the horizontal axis (or time axis) position where the velocity value reaches its minimum value within a specific local neighborhood.

[0053] In some embodiments, since cardiac activity is a continuous process, the velocity information related to the heart can exhibit a certain periodicity in the ultrasound spectral image. Therefore, the minimum value location can correspond to a key node in cardiac activity, and the portion between two adjacent minimum value locations is very likely to contain a complete blood flow peak. Therefore, a single-peak sub-image can be divided by iteratively traversing multiple minimum value locations.

[0054] In some embodiments, after obtaining the intermediate curve, the processing device 120 may employ an image recognition algorithm to determine multiple minimum value locations. Exemplary image recognition algorithms may include extreme value detection algorithms, template matching algorithms, thresholding algorithms, etc. For example, the processing device 120 may compare the summation values ​​of adjacent pixels on the intermediate curve to identify a specific point where the velocity value reaches a local minimum. The processing device 120 may determine the location of this specific point as the minimum value location. The processing device 120 may crop the intermediate ultrasound spectrum image along the minimum value location to obtain single-peak sub-images of each single peak.

[0055] Furthermore, the processing device 120 can calculate the start timestamp of each single-peak sub-image based on the start timestamp of the current frame of ultrasound spectrum image (i.e., the timestamps corresponding to the leftmost and rightmost sides).

[0056] For example, suppose the current frame ultrasound spectrum image is the first... Frame of ultrasound spectral image (i.e., the frame calculated in this case is the first frame) (calculated once), the timestamps corresponding to its leftmost and rightmost edges are respectively and Then the time interval corresponding to one pixel in the horizontal axis direction It can be determined by the following formula (1): (1) in, For the first The width of the ultrasound spectral image frame. Then the... Zhang Danfeng's sub-plot start timestamp It can be determined by the following formula (2): (2) in, For the first The x-coordinate of the left boundary of Zhang Danfeng's subgraph.

[0057] In step 440, the processing device 120 can determine whether a single-peak subgraph is a subgraph to be processed based on the start timestamp of each single-peak subgraph. In some embodiments, the processing device 120 can determine all single-peak subgraphs corresponding to the intermediate ultrasound spectrum image as subgraphs to be processed.

[0058] In some embodiments, to filter out old blood flow peak sub-images that have already been calculated in the previous frame, avoid data redundancy and repeated calculations caused by the rolling spectrum image, and ensure the efficient and smooth operation of the real-time system, the processing device 120 can determine single-peak sub-images with a start timestamp greater than a reference timestamp as sub-images to be processed. The reference timestamp is the peak timestamp corresponding to the last complete blood flow peak in the previous frame of the ultrasound spectrum image. For example, if the first... The peak position of the last complete blood flow peak corresponding to the frame ultrasound spectral image is (i.e., there are currently M complete blood flow peaks), then the processing device 120 can process peaks with a start timestamp greater than... The unimodal subgraph is identified as the subgraph to be processed.

[0059] In some embodiments, the processing device 120 may employ a trained machine learning model to determine multiple single-peak subplots. As an example only, the processing device 120 may input an intermediate ultrasound spectrum image into the trained machine learning model. The trained machine learning model may then output multiple single-peak subplots. The trained machine learning model can be obtained by training multiple training samples. Each training sample may include a sample intermediate ultrasound spectrum image, which is the portion of the sample ultrasound spectrum image below the horizontal axis. Each training sample may also include multiple sample single-peak subplots corresponding to the sample ultrasound spectrum image. When training the machine learning model, the multiple sample single-peak subplots corresponding to the sample ultrasound spectrum image can be used as labels.

[0060] According to some embodiments in this specification, the location of blood flow peaks can be quickly located through binarization and envelope extraction. Furthermore, by using the "reference timestamp (the peak value of the last complete blood flow peak in the previous frame of ultrasound spectrum image)" as the truncation standard, old blood flow peak images that have already been calculated in the previous frame are effectively filtered out. This avoids data redundancy and repeated calculations caused by rolling spectrum images, ensuring the real-time, efficient, and smooth operation of the system.

[0061] In step 330, the processing device 120 can input each sub-image to be processed into the evaluation and mask determination model to obtain a quality score and mask image for each sub-image to be processed. In some embodiments, step 330 can be performed by the target sub-image determination module 230 in the system 200.

[0062] A quality score reflects the completeness of blood flow peaks in the sub-image to be processed. A higher quality score indicates greater completeness of the blood flow peaks in the sub-image. A mask image is used to extract the blood flow peaks from the corresponding sub-image. In some embodiments, the mask image includes pixels with specified values ​​(e.g., 0 or 1). For example, the mask image may be a binary image, including pixels corresponding to regions of interest (ROIs) with a value of "1" and pixels corresponding to regions outside the ROI with a value of "0".

[0063] By evaluating and determining the model simultaneously to obtain the quality score and mask image of each sub-image to be processed, resource consumption can be reduced and system performance can be improved. At the same time, the tedious steps of manual measurement can be avoided, subjective errors can be reduced, thereby improving detection efficiency, improving the accuracy of blood ejection peak position determination, and realizing the automation of ultrasound spectrum image related parameter measurement.

[0064] In some embodiments, the evaluation and mask determination model is a machine learning model. For example, the evaluation and mask determination model includes one or more of the following: deep neural network (DNN) models, neural network models, or recurrent neural network models.

[0065] In some embodiments, the evaluation and mask determination model may be pre-trained by, for example, a device or system other than the ultrasound imaging system 100 (e.g., a vendor-made device or system). For example, the processing device 120 may train the evaluation and mask determination model using machine learning algorithms (e.g., support vector machine algorithms) or deep learning algorithms (e.g., convolutional neural networks) based on multiple training samples and their corresponding labels. The training samples may include multiple sample sub-images to be processed, each bearing both a first label and a second label. The first label includes a reference mask image of the blood flow peaks generated based on the sample sub-images to be processed. The second label includes a reference quality score for complete or incomplete blood flow peaks in the sample sub-images to be processed. It should be noted that when a peak in a sample sub-image is not a blood flow peak, its quality score may be set to 0. In this case, the first label for the sample sub-image may be empty. It should be noted that the first and second labels may be manually annotated by experienced physicians.

[0066] In some embodiments, the processing device 120 acquires a large number of training samples and their corresponding labels as a training dataset, and performs multiple iterations. When the iteration termination condition is met, the iteration ends, and a trained model is obtained. At least one iteration in the multiple iterations includes: selecting one or more training samples from the training dataset as input to the evaluation and mask determination model to obtain the model prediction output corresponding to one or more training samples; substituting the model prediction output and the corresponding labels into a predefined loss function formula to calculate the value of the loss function; and updating the model parameters in the model in reverse based on the value of the loss function. Reverse updating can be implemented in various ways, such as updating using gradient descent. The iteration condition can be the convergence of the loss function, the number of iterations reaching a threshold, etc.

[0067] In some embodiments, the evaluation and mask determination model may include an encoding layer and a decoding layer. The encoding layer uses MobileNetV4-ConvSmall as the backbone network for feature extraction from the sub-image to be processed. The decoding layer includes a segmentation decoding layer and a quality assessment decoding layer. The segmentation decoding layer uses ResUNet for feature decoding, receiving multi-scale features from the encoding layer and fusing them to output a mask image of the blood flow peaks. The quality assessment decoding layer selects the output of the last layer of the encoding layer as input, performs average pooling and flattening, and then inputs it into a Linear layer, finally outputting a quality score. Based on the blood flow peak contours and quality scores in the pre-labeled sample sub-images to be processed, the model is trained. During training, the two labels are subjected to a cross-entropy Dice mixed loss and a Huber Loss loss, respectively, and the sum of the two losses is calculated as the total loss. The AdamW optimizer is used for model training. After the model training converges, the model that performs best on the validation set is selected as the evaluation and mask determination model. In application, the subgraph to be processed is input into the evaluation and mask determination model. After processing by the encoding layer, it passes through the decoding layer (i.e., the segmentation decoding layer and the quality evaluation decoding layer) to output the corresponding prediction results. More details about the evaluation and mask determination model can be found in other parts of this specification, such as... Figure 5 Its description.

[0068] In some embodiments, the processing device 120 may preprocess the subgraph to be processed before inputting it into the evaluation and mask determination model. In some embodiments, preprocessing may include operations such as denoising and normalization to ensure the quality of the input data.

[0069] According to some embodiments of this specification, by using a machine learning model for blood flow peak contour segmentation and quality scoring, it is possible to automatically and accurately remove single-peak sub-images corresponding to incomplete peaks, no blood flow peaks, or blood flow peaks in non-ejection phases, ensuring that the blood flow peaks involved in the subsequent calculation of target parameters are high-quality blood flow peaks with complete features, thereby significantly improving the accuracy and clinical reference value of the final measurement parameters.

[0070] In step 340, the processing device 120 may select one or more target sub-images from the multiple sub-images to be processed based on their quality scores. In some embodiments, step 340 may be performed by the target sub-image determination module 230 in the system 200. A target sub-image refers to a sub-image to be processed that contains a complete blood flow peak.

[0071] In some embodiments, the processing device 120 may identify sub-images with a quality score greater than or equal to a scoring threshold as target images. This allows sub-images with complete blood flow peaks to be retained, while sub-images containing only incomplete blood flow peaks or non-ejection phase blood flow peaks to be removed. The scoring threshold may be set according to the default settings of the ultrasound imaging system 100 or by the operator via the terminal device 140.

[0072] In step 350, the processing device 120 may extract the contour of each target sub-image based on a mask image of one or more target sub-images to obtain the contour of the blood flow peak in each target sub-image. In some embodiments, step 350 may be performed by the target parameter determination module 240 in the system 200.

[0073] For example, the processing device 120 can determine the contour of the blood flow peak by multiplying the mask image with the corresponding target sub-image. Specifically, the processing device 120 can determine the contour of the blood flow peak by multiplying the value of each pixel in the target sub-image with the value of the corresponding pixel in the mask image.

[0074] In step 360, the processing device 120 may calculate target parameters based on the contours of blood flow peaks in one or more target sub-images and display the target parameters to the user. In some embodiments, step 360 may be performed by the target parameter determination module 240 in system 200.

[0075] In some embodiments, the target parameters include any one or more combinations of the above indicators such as maximum velocity (Vmax), velocity time integral (VTI), stroke volume (SV), the start time stamp, end time stamp, peak time stamp of the blood flow peak, heart rate (HR) and cardiac output (CO) corresponding to the current frame of ultrasound spectral image.

[0076] As an example only, for each target sub-image, the processing device 120 can determine the velocity corresponding to the extreme point of the blood flow peak contour in the target sub-image as the corresponding blood flow peak velocity. The processing device 120 can calculate the area of ​​the blood flow peak profile, and by multiplying it by the scale of the horizontal and vertical axes, the corresponding time-velocity integral can be obtained. Furthermore, stroke volume It can be calculated using the following formula (3): (3) in, The diameter of the left ventricular outflow tract can be manually entered by a physician or obtained from system measurements. In some embodiments, the processing device 120 can calculate the current number of... Average parameters calculated in this cycle: (4) (5) (6) in, This represents the average peak blood flow velocity corresponding to the previous frame of the ultrasound spectrum image. This represents the average value of the time-velocity integral corresponding to the previous frame of the ultrasound spectrum image. This represents the average stroke volume corresponding to the previous frame of the ultrasound spectrum image. This represents the number of complete blood flow peaks in the current frame of the ultrasound spectral image, which is also the number of cardiac cycles.

[0077] In some embodiments, the processing device 120 may also calculate the timestamp corresponding to the newly added blood flow peak value through the following (7): (7) in, For the newly added The horizontal axis coordinates corresponding to the peak values ​​of each peak are calculated, along with the left and right start and end timestamps for each peak. and Then the current number heart rate Heart output It can be calculated using (8) and (9) below respectively: (8) (9) In some embodiments, the processing device 120 can map the blood flow peak segmentation results, quality scores, and / or target parameters back onto the current frame ultrasound spectrogram and display them to the user (e.g., a doctor), such as... Figure 6 As shown. In some embodiments, a user (e.g., an experienced doctor) can adjust one or more target parameters via a terminal device (e.g., terminal device 140).

[0078] According to some embodiments of this specification, by mapping blood flow peak segmentation results, quality scores, and / or target parameters to the current frame ultrasound spectral image and displaying them to the user (e.g., a physician), the user can perform the following freeze operation based on the displayed relevant parameters and / or segmentation results to find a satisfactory waveform, rather than relying solely on experience. This avoids subjective errors in human interpretation, thereby improving the accuracy of analysis and supporting automated diagnosis and standardized assessment.

[0079] In some embodiments, in response to a user performing a freeze operation, the processing device 120 may also determine a target ultrasound spectral image and display it to the user based on the start timestamps of all blood flow peaks in all acquired target sub-images and their corresponding quality scores. In some embodiments, step 370 may be performed by the target parameter determination module 240 in system 200.

[0080] A freeze operation refers to an action triggered by a user (e.g., a physician) during an ultrasound examination, where the ultrasound imaging system 100 stops updating the ultrasound spectral image in real time and locks and calculates the final cardiac hemodynamic parameters. The user can determine whether a freeze operation is necessary based on the target parameters of all blood flow peaks in all acquired target sub-images.

[0081] In response to a user-initiated freeze operation, processing device 120 stops updating the current frame of the ultrasound spectrogram. After completing the current calculation, it obtains all blood flow peaks that meet the requirements across the entire scrolling ultrasound spectrogram. At this point, processing device 120 can sort all blood flow peaks in all acquired target sub-images in chronological order. Then, processing device 120 can use a time window to frame all blood flow peaks and determine the target blood flow peak within each frame of signal. The window length of the time window... The time length corresponds to the width of a single frame of ultrasound spectral image. The target blood flow peak is the blood flow peak completely included within the frame signal. Further, the processing device 120 can calculate the overall quality score of the target blood flow peak. The overall quality score is the sum of the quality scores of all target sub-images corresponding to the target blood flow peaks within the frame signal. Finally, the processing device 120 can determine the frame signal with the highest overall quality score as the target ultrasound spectral image and display it to the user. Further, the processing device 120 can update the target parameters based on the target parameters corresponding to the target blood flow peaks within the target ultrasound spectral image and display them to the user.

[0082] In some embodiments, since the heart rate calculated over a longer time span is more accurate, the processing device 120 can... The average heart rate obtained from the calculation: The heart rate of the final subject being examined is used as the basis for determining cardiac output. ,Right now: (10) In some embodiments, in addition to the heart rate described above Cardiac output In addition, the processing device 120 can use other target parameters corresponding to the target ultrasound spectrum image as the final parameters displayed to the user. In some embodiments, the user (e.g., a doctor) can manually select other peaks or the average of several other peaks as the final display result as needed. Furthermore, the user can perform diagnostic assessments of the examinee's cardiac function based on the displayed target ultrasound spectrum image and its corresponding parameters.

[0083] According to some embodiments of this specification, by automatically calculating and displaying parameters in the background in real time before freezing, doctors can accurately grasp the timing of freezing; after freezing, the system can automatically locate and display the optimal ultrasound spectrum image based on the comprehensive quality score, thereby greatly improving the automation level of the examination, measurement efficiency, and the objectivity and validity of the measurement results.

[0084] It should be noted that the above description is for convenience only and should not limit this application to the scope of the embodiments described. It is understood that those skilled in the art, after understanding the principle of the system, can make various modifications and changes in form and detail to the application fields of the above methods and systems without departing from this principle.

[0085] Figure 5 This is a schematic diagram illustrating the use of an exemplary evaluation and mask determination model 500 to determine the quality score and mask image for each sub-image to be processed, according to some embodiments of this specification. The input to the evaluation and mask determination model 500 may include the sub-image to be processed, and the output may include the quality score and mask image of the sub-image to be processed. Figure 5 As shown, the evaluation and mask determination model 500 may include an encoding layer 510, a segmentation decoding layer 520, and a quality evaluation decoding layer 530. Both the segmentation decoding layer 520 and the quality evaluation decoding layer 530 can be connected to the encoding layer 510.

[0086] The coding layer 510 can be configured to extract feature maps from the input subgraph to be processed. The input to the coding layer 510 may include the subgraph to be processed, and the output may include the feature maps of the subgraph to be processed. In some embodiments, the coding layer 510 may employ MobileNetV4. The ConvSmall structure serves as the backbone network for feature extraction from the subgraph to be processed. In some embodiments, the coding layer 510 may include multiple modules, such as Block 1, Block 2, Block 3, Block 4, Block 5, and Block 6.

[0087] By way of example only, Block 1 may include a convolutional layer (convbn). The input to Block 1 is the subgraph to be processed, and the output is the first feature map of the subgraph to be processed. In this specification, a convolutional layer may consist of a 2D convolutional layer (conv2d), a BatchNorm2d layer, and a ReLU layer. Block 2 may include two convolutional layers (convbn). The input to Block 2 is the first feature map, and the output is the second feature map of the subgraph to be processed. Block 3 may include two convolutional layers (convbn). The input to Block 3 is the second feature map, and the output is the third feature map of the subgraph to be processed. Block 4 may include six uib layers, such as first uib layer 1, second uib layer 1, third uib layer 1, fourth uib layer 1, fifth uib layer 1, and sixth uib layer 1. In some embodiments, each uib layer may include multiple convolutional layers, such as 3, 4, 5, etc. For example, the first uib layer 1 may include four convolutional layers (convbn). Block 4 can include three convbn layers for each of its 1st, 2nd, 3rd, 4th, 5th, and 6th UIB layers. The input to Block 4 is the third feature map, and the output is the fourth feature map of the sub-image to be processed. Block 5 can include six UIB layers, such as 2nd, 2nd, 3rd, 2nd, 4th, 5th, and 6th UIB layers. Both 2nd and 2nd UIB layers can include four convbn layers. The 2nd, 3rd, 4th, 5th, and 6th UIB layers can each include three convbn layers. The input to Block 5 is the fourth feature map, and the output is the fifth feature map of the sub-image to be processed. Block 6 can include two convbn layers. The input to Block 6 is the fifth feature map, and the output is the sixth feature map of the sub-image to be processed.

[0088] The segmentation decoding layer 520 can be configured to generate a mask image of the sub-image to be processed. The input to the segmentation decoding layer 520 may include feature maps of the sub-image to be processed, and the output may include a mask image of the sub-image to be processed. In some embodiments, the segmentation decoding layer 520 may employ a ResNet network for feature decoding. In some embodiments, the segmentation decoding layer 520 may include multiple modules, such as Block 11, Block 22, Block 33, Block 44, and a Head layer. As an example only, Block 11, Block 22, Block 33, and Block 44 may each include an attention module and a ResNet block, and the Head layer may include three convbn layers. The input to Block 11 is a fourth feature map and a sixth feature map, and the output is a first fused image. The input to Block 22 is a third feature map and a first fused image, and the output is a second fused image. The input to Block 33 is a second feature map and a second fused image, and the output is a third fused image. The input to Block 44 is a first feature map and a third fused image, and the output is a fourth fused image. The input to the Head layer is the fourth fusion map, and the output is a mask image used to extract blood flow peaks from the sub-image to be processed.

[0089] The quality assessment decoding layer 530 can be configured to generate a quality score for the sub-image to be processed. The input to the quality assessment decoding layer 530 may include a feature map of the sub-image to be processed, and the output may include a quality score for the sub-image to be processed. In some embodiments, the quality assessment decoding layer 530 may sequentially include an average pooling layer, a flattening layer, a linear layer, a ReLU layer, and another linear layer. The input to the quality assessment decoding layer 530 is a sixth feature map, and the output is a quality score used to assess the integrity of the blood flow peaks in the sub-image to be processed.

[0090] After the segmentation decoding layer 520 outputs the mask image, the mask image can be applied to the corresponding sub-image to be processed to obtain the blood flow peak contour. Then, the blood flow peak contour and quality score can be mapped back to the corresponding ultrasound spectrum image and displayed to the user.

[0091] Figure 6 This is a schematic diagram illustrating the output results of a single calculation according to some embodiments of this specification. For example... Figure 6 As shown, one or more target parameters are displayed on the current frame of the ultrasound spectral image. These target parameters include a quality score corresponding to each target sub-image (e.g., Figure 6 middle ), the start timestamp of the blood flow peak (e.g., Figure 6 middle , , ), end timestamp (e.g., Figure 6 middle , , ), peak timestamp (e.g., Figure 6 middle , , Heart rate (HR), cardiac output (CO), quality score, blood flow peak profile, or any combination thereof.

[0092] According to some embodiments of this specification, measurements can be completed automatically without the need for doctors to manually click on peak positions and time intervals, significantly improving the efficiency of doctor examinations and enhancing the objectivity of measurement results. Furthermore, unlike manual measurements that typically provide only one blood flow peak measurement result, the target parameter determination method provided in some embodiments of this specification can generate measurement results for all blood flow peaks in the current frame of the ultrasound spectral image, making the final measurement value more reasonable.

[0093] Figure 7A This is an exemplary flowchart illustrating the determination of a subgraph to be processed according to some embodiments of this specification. Figure 7B This is a schematic diagram of the quality scoring of the subgraph to be processed according to some embodiments of this specification. Figure 7C This is a schematic diagram showing the results of sorting all blood flow peaks by time according to some embodiments of this specification. Figure 7A As shown, the ultrasound spectrum image 710 can be segmented into an intermediate ultrasound spectrum image 720 based on the horizontal axis of the cardiac ultrasound spectrum image 710. The intermediate ultrasound spectrum image 720 can be binarized to generate a binarized intermediate ultrasound spectrum image 730. Simultaneously, the envelope of the intermediate ultrasound spectrum image 720 can be extracted to obtain an intermediate ultrasound spectrum image 740 with the extracted envelope. The vertical axis of the envelope is extracted and smoothed to finally obtain an intermediate curve 750. Multiple minimum value positions A in the intermediate curve 750 can be identified. Based on these multiple minimum value positions A, the intermediate ultrasound spectrum image 720 can be divided into multiple candidate sub-images 760 to be processed. For example, the portion between any two minimum value positions A in the intermediate ultrasound spectrum image 720, or the portion between the timestamp corresponding to the minimum value position A and the timestamp corresponding to the start or end position of the horizontal axis of the intermediate ultrasound spectrum image, can be determined as candidate sub-images to be processed 760. Candidate sub-images 760 whose start timestamp is after the peak value of the last complete blood flow peak in the previous frame of ultrasound spectrum image can be identified as sub-images 770 to be processed.

[0094] Furthermore, multiple sub-images 770 to be processed can be input into the evaluation and mask determination model to obtain the quality score and mask image of each sub-image 770. For example, for each sub-image 770 to be processed, after inputting it into the evaluation and mask determination model, the model can output the quality score and corresponding mask image of the sub-image 770. One or more target sub-images can be determined from the multiple sub-images 770 based on their quality scores. For example, if the quality score of a sub-image is greater than or equal to a scoring threshold, it can be determined as a target sub-image. If the quality score of a sub-image is less than the scoring threshold, it can be removed. For example, if... Figure 7B As shown, when the scoring threshold is 3, sub-image 772 with a quality score greater than or equal to 3 can be identified as the target sub-image. Sub-images 774, 776, and 778 with a quality score less than 3 can be discarded. Further, target parameters for the target sub-images can be calculated. These target parameters can then be mapped back onto the ultrasound spectral image 710 for display to the user.

[0095] In some embodiments, such as Figure 7C As shown, by sorting all blood flow peaks in all acquired target sub-images in chronological order and using the time length corresponding to the width of a single frame of ultrasound spectral image, Using a time window, iterate through existing blood flow peaks, with each iteration starting from the first peak. Onset time of each blood flow peak Starting from, As the endpoint, calculate the sum of the quality scores of all blood flow peaks included within the time window, and use this sum as the overall quality score for the current window.

[0096] Optionally, the processing device 120 can select the window with the highest overall quality score as the final display window (i.e., the target ultrasound spectrum image). Optionally, since the heart rate value calculated over a longer time span is more accurate, the average heart rate corresponding to all ultrasound spectrum images is displayed as the target parameter. The remaining display parameters can be the average values ​​of parameters calculated based on each peak within the target ultrasound spectrum image. Optionally, the physician can manually select other peaks or the average values ​​of several other peaks as the final result as needed.

[0097] The beneficial effects of some embodiments in this specification include, but are not limited to, the following aspects: (1) The image processing technology automatically identifies the position of blood flow peaks, eliminating the need for doctors to manually click on the peak position and outline the contour, greatly reducing the difficulty of doctors' operation, improving examination efficiency, and avoiding subjective errors of doctors, thereby improving the objectivity and consistency of measurement results; (2) The deep learning model is used for blood flow peak contour segmentation and quality scoring, with high contour extraction accuracy and accurate target sub-image selection, which can effectively distinguish between complete and incomplete blood flow peaks, ensuring the accuracy of target parameters; (3) Before the doctor performs the freeze operation, the system automatically calculates each blood flow peak in real time. The relevant parameters (Vmax, VTI, SV, CO, etc.) are displayed to the doctor, who then performs the freeze operation based on the displayed parameters to find a satisfactory waveform, rather than relying solely on experience; (4) After the doctor performs the freeze operation, the system automatically finds the location of the optimal ultrasound spectrum based on the quality scores of multiple blood flow peaks in each frame of the ultrasound spectrum, eliminating the need for the doctor to manually drag the spectrum for selection; (5) It eliminates the need to rely on ECG information to assist in determining the cardiac cycle, freeing the doctor from dependence on ECG and eliminating the need for the doctor to attach ECG electrodes to the patient before each examination, saving time and costs and better conforming to the operating habits of clinicians. It should be understood that the above beneficial effects can be reflected in different embodiments of this specification, and the same embodiment may also have one or more of the aforementioned beneficial effects at the same time.

[0098] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore remain within the spirit and scope of the exemplary embodiments of this application.

Claims

1. An automatic measurement method for cardiac Doppler ultrasound spectral images, characterized in that, include: According to a preset time interval, the current frame ultrasound spectrum image is continuously updated, and the current frame ultrasound spectrum image is processed to obtain one or more sub-images of blood flow peaks to be processed. Each sub-image to be processed is input into the evaluation and mask determination model to obtain a quality score and mask image for each sub-image to be processed. Based on the quality scores of the one or more sub-images to be processed, one or more target sub-images are selected from the one or more sub-images to be processed. Based on the mask images of one or more target sub-images, contour extraction is performed on each target sub-image to obtain the contour of the blood flow peak in each target sub-image. as well as Based on the contours of blood flow peaks in one or more target sub-images, target parameters are calculated and displayed to the user.

2. The method according to claim 1, characterized in that, The method further includes: In response to the user performing a freeze operation, a target ultrasound spectral image is determined and displayed to the user based on the start timestamps of all blood flow peaks in all acquired target sub-images and their corresponding quality scores. The user confirms whether the freeze operation needs to be performed based on the target parameters of all blood flow peaks in all acquired target sub-images.

3. The method according to claim 1, characterized in that, The process of processing the current frame ultrasound spectral image to obtain one or more sub-images of blood flow peaks includes: Based on the horizontal axis of the current frame ultrasound spectrum image, the current frame ultrasound spectrum image is segmented to obtain the middle ultrasound spectrum image below the horizontal axis. The intermediate ultrasound spectrum image is binarized and its envelope is extracted to determine the intermediate curve and smooth it. Based on the multiple minimum locations in the intermediate curve, the intermediate ultrasound spectrum image is cropped to obtain single-peak sub-images for each single peak, and the start timestamp of each single-peak sub-image is calculated; and Based on the start timestamp of each unimodal subgraph, determine whether the unimodal subgraph is a subgraph to be processed.

4. The method according to claim 3, characterized in that, The step of determining whether a single-peaked subgraph is a subgraph to be processed based on the start timestamp of each single-peaked subgraph includes: A single-peak sub-image with a start timestamp greater than a reference timestamp is identified as a sub-image to be processed. The reference timestamp is the peak timestamp of the blood flow peak in the last target sub-image corresponding to the previous frame of ultrasound spectrum image.

5. The method according to claim 1, characterized in that, The step of selecting one or more target sub-images from the one or more sub-images to be processed based on their quality scores includes: The subgraphs to be processed that have a quality score greater than or equal to the score threshold are identified as the target subgraphs.

6. The method according to claim 2, characterized in that, In response to the user performing a freeze operation, the system determines the target ultrasound spectral image and displays it to the user based on the start timestamps of all blood flow peaks in all acquired target sub-images and their corresponding quality scores, including: In response to the user performing the freeze operation, the updating of the current frame ultrasound spectrum image is stopped; Sort all blood flow peaks in all acquired target sub-maps in chronological order, and perform the following operations on all blood flow peaks sequentially: Using a time window, all blood flow peaks are divided into frames, and the window length of the time window is the time length corresponding to the width of a single frame of ultrasound spectrum image; Determine the target blood flow peak within each frame of signal, wherein the target blood flow peak is a blood flow peak completely included within the frame of signal; Calculate the overall quality score of the target blood flow peak, where the overall quality score is the sum of the quality scores of the target sub-images corresponding to all target blood flow peaks within the frame signal; and The frame with the highest overall quality score is identified as the target ultrasound spectrum image.

7. The method according to claim 6, characterized in that, After determining the target ultrasound spectral image, the method further includes: The target parameters are updated based on the target parameters corresponding to the target blood flow peaks within the target ultrasound spectrum image.

8. The method according to claim 1, characterized in that, The method further includes: The preset time interval is updated based on the calculated cardiac cycle of the subject.

9. The method according to claim 1, characterized in that, The target parameters include at least one of the following: peak velocity Vmax, time-velocity integral VTI, stroke volume SV, peak start time stamp, peak end time stamp, peak time stamp, heart rate HR, cardiac output CO, quality score, and peak profile.

10. An automatic measurement system for cardiac Doppler ultrasound spectral images, comprising a processor, characterized in that, The processor is used to execute the automatic measurement method for cardiac Doppler ultrasound spectral images according to any one of claims 1-9.

11. A computer-readable storage medium storing computer instructions, wherein when a computer reads the computer instructions in the storage medium, the computer executes the automatic measurement method for cardiac Doppler ultrasound spectral images as described in any one of claims 1-9.

Images