Methods and devices for umbilical blood flow ultrasound image measurement and parallel processing

Abstract

This application relates to a method and apparatus for measuring and parallel processing umbilical blood flow ultrasound images. The method includes: accelerating the acquisition of a standard cross-section using data parallelism and convolution acceleration; identifying the region of interest (ROI), X-axis region, and umbilical blood flow spectral envelope of the standard cross-section using digital image processing technology, thereby calculating the peaks and troughs of the umbilical blood flow spectrum and locating a continuous and stable umbilical blood flow spectrum. Further identification of scale points allows for the rapid location of velocity and time scales, enabling efficient calculation of the conversion ratios of the Y-axis velocity scale and the X-axis time scale. Based on this, the calculated continuous and stable umbilical blood flow spectrum, the conversion ratio of the Y-axis velocity scale, and the conversion ratio of the X-axis time scale are used to automatically and accurately measure the correlation coefficient of the umbilical blood flow spectrum. This process requires no manual intervention, significantly improving the calculation efficiency of the umbilical blood flow spectrum correlation coefficient.

Description

Technical Field

[0001] This application relates to the fields of ultrasound technology and image processing technology, and in particular to a method, apparatus and computer equipment for measuring and parallel processing ultrasound images of umbilical blood flow. Background Technology

[0002] With the deepening integration of the medical and computer fields, deep learning and digital image processing technologies are widely used in medical technology, bringing convenience to routine medical diagnoses for healthcare professionals. The detection of umbilical cord blood flow in ultrasound images and the measurement of related growth parameters are of great significance for predicting issues such as fetal growth restriction and fetal hypoxia.

[0003] Currently, clinicians rely heavily on medical instruments to measure growth parameters related to umbilical blood flow. Measuring umbilical blood flow images without medical instruments requires manually selecting the umbilical blood flow spectral envelope, measuring velocity, time scale points, calculating the spectral envelope area, and selecting and calculating spectral peaks and troughs before measurement can be performed. The operation is complex, time-consuming, and labor-intensive. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for umbilical blood flow ultrasound image measurement and parallel processing that can improve processing efficiency in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a method for measuring and parallel processing umbilical blood flow ultrasound images. The method includes:

[0006] Accelerate the acquisition of standard cross-sections of umbilical blood flow through GPU parallel computing;

[0007] Locate the region of interest for umbilical blood flow in the standard section of the umbilical blood flow;

[0008] Determine the X-axis region from the region of interest of the umbilical blood flow;

[0009] The umbilical blood flow spectral envelope is determined based on the region of interest of the umbilical blood flow and the X-axis region;

[0010] Calculate the peak and trough points of the umbilical blood flow spectrum based on the umbilical blood flow spectrum envelope;

[0011] Based on the peaks and troughs of the umbilical blood flow spectrum, a continuous and stable umbilical blood flow spectrum is selected.

[0012] Locate the first region of interest (ROI) of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first ROI.

[0013] Locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the region of interest;

[0014] The correlation coefficient of the umbilical blood flow spectrum is obtained based on the conversion ratio of the continuous and stable umbilical blood flow spectrum, the velocity scale in the Y-axis direction, and the time scale in the X-axis direction.

[0015] Secondly, this application also provides a device for measuring and parallel processing umbilical blood flow ultrasound images. The device includes:

[0016] The section acquisition module is used to accelerate the acquisition of standard sections of umbilical blood flow through GPU parallel computing;

[0017] The region of interest (ROI) localization module is used to locate the ROI of the umbilical blood flow in the standard cross-section of the umbilical blood flow.

[0018] The X-axis analysis module is used to determine the X-axis region from the region of interest of the umbilical blood flow.

[0019] The envelope analysis module is used to determine the umbilical blood flow spectrum envelope based on the region of interest of the umbilical blood flow and the X-axis region;

[0020] The first calculation module is used to calculate the peak and trough points of the umbilical blood flow spectrum based on the envelope of the umbilical blood flow spectrum.

[0021] The spectrum selection module is used to select a continuous and stable umbilical blood flow spectrum based on the peak and trough points of the umbilical blood flow spectrum.

[0022] The first conversion ratio calculation module is used to locate the first region of interest of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first region of interest.

[0023] The second conversion ratio calculation module is used to locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the region of interest.

[0024] The coefficient calculation module is used to obtain the correlation coefficient of the umbilical blood flow spectrum based on the continuous and stable umbilical blood flow spectrum, the conversion ratio of the velocity scale in the Y-axis direction, and the conversion ratio of the time scale in the X-axis direction.

[0025] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0026] Accelerate the acquisition of standard cross-sections of umbilical blood flow through GPU parallel computing;

[0027] Locate the region of interest for umbilical blood flow in the standard section of the umbilical blood flow;

[0028] Determine the X-axis region from the region of interest of the umbilical blood flow;

[0029] The umbilical blood flow spectral envelope is determined based on the region of interest of the umbilical blood flow and the X-axis region;

[0030] Calculate the peak and trough points of the umbilical blood flow spectrum based on the umbilical blood flow spectrum envelope;

[0031] Based on the peaks and troughs of the umbilical blood flow spectrum, a continuous and stable umbilical blood flow spectrum is selected.

[0032] Locate the first region of interest (ROI) of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first ROI.

[0033] Locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the region of interest;

[0034] The correlation coefficient of the umbilical blood flow spectrum is obtained based on the conversion ratio of the continuous and stable umbilical blood flow spectrum, the velocity scale in the Y-axis direction, and the time scale in the X-axis direction.

[0035] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps:

[0036] Accelerate the acquisition of standard cross-sections of umbilical blood flow through GPU parallel computing;

[0037] Locate the region of interest for umbilical blood flow in the standard section of the umbilical blood flow;

[0038] Determine the X-axis region from the region of interest of the umbilical blood flow;

[0039] The umbilical blood flow spectral envelope is determined based on the region of interest of the umbilical blood flow and the X-axis region;

[0040] Calculate the peak and trough points of the umbilical blood flow spectrum based on the umbilical blood flow spectrum envelope;

[0041] Based on the peaks and troughs of the umbilical blood flow spectrum, a continuous and stable umbilical blood flow spectrum is selected.

[0042] Locate the first region of interest (ROI) of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first ROI.

[0043] Locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the region of interest;

[0044] The correlation coefficient of the umbilical blood flow spectrum is obtained based on the conversion ratio of the continuous and stable umbilical blood flow spectrum, the velocity scale in the Y-axis direction, and the time scale in the X-axis direction.

[0045] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps:

[0046] Standard cross-sections of umbilical blood flow are obtained through parallel computing using GPUs.

[0047] Locate the region of interest for umbilical blood flow in the standard section of the umbilical blood flow;

[0048] Determine the X-axis region from the region of interest of the umbilical blood flow;

[0049] The umbilical blood flow spectral envelope is determined based on the region of interest of the umbilical blood flow and the X-axis region;

[0050] Calculate the peak and trough points of the umbilical blood flow spectrum based on the umbilical blood flow spectrum envelope;

[0051] Based on the peaks and troughs of the umbilical blood flow spectrum, a continuous and stable umbilical blood flow spectrum is selected.

[0052] Locate the first region of interest (ROI) of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first ROI.

[0053] Locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the region of interest;

[0054] The correlation coefficient of the umbilical blood flow spectrum is obtained based on the conversion ratio of the continuous and stable umbilical blood flow spectrum, the velocity scale in the Y-axis direction, and the time scale in the X-axis direction.

[0055] The aforementioned method, apparatus, computer equipment, storage medium, and computer program products for umbilical blood flow ultrasound image measurement and parallel processing fully utilize the computing power of multiple GPU processing units, distributing tasks across multiple units for parallel processing. This accelerates the acquisition of standard umbilical blood flow sections. By employing digital image processing technology to identify the region of interest (ROI), X-axis region, and umbilical blood flow spectral envelope within the standard umbilical blood flow section, the peaks and troughs of the umbilical blood flow spectrum can be calculated, allowing the location of continuous and stable umbilical blood flow spectra. Further identification of scale points enables rapid location of velocity and time scales, thereby efficiently calculating the conversion ratios of the Y-axis velocity scale and the X-axis time scale. Based on this, using the calculated continuous and stable umbilical blood flow spectrum, the conversion ratios of the Y-axis velocity scale and the X-axis time scale, the correlation coefficient of the umbilical blood flow spectrum can be automatically and accurately measured. This process requires no manual intervention, significantly improving the computational efficiency of the umbilical blood flow spectrum correlation coefficient. Attached Figure Description

[0056] Figure 1 This is a diagram illustrating the application environment of a method for measuring and parallel processing umbilical blood flow ultrasound images in one embodiment.

[0057] Figure 2 This is a flowchart illustrating the method for measuring and parallel processing umbilical blood flow ultrasound images in one embodiment;

[0058] Figure 3 This is a schematic diagram of a standard cross-section of umbilical blood flow in one embodiment;

[0059] Figure 4 This is a schematic diagram of a binarized image of a standard section of umbilical blood flow in one embodiment;

[0060] Figure 5 This is a segmentation diagram of the region of interest for umbilical blood flow in one embodiment;

[0061] Figure 6 This is a schematic diagram of the ROI region of umbilical blood flow in one embodiment;

[0062] Figure 7 This is an example of an edge feature map along the X-axis.

[0063] Figure 8 Here is an X-ray diagram of umbilical blood flow in one embodiment;

[0064] Figure 9 Here is a binarized map of umbilical blood flow ROI in one embodiment;

[0065] Figure 10 An example of an umbilical blood flow spectrum envelope diagram

[0066] Figure 11This is a schematic diagram of the umbilical blood flow spectrum offset in one embodiment;

[0067] Figure 12 This is a peak and trough diagram of umbilical blood flow spectrum in one embodiment;

[0068] Figure 13 Here is a continuous stable and umbilical blood flow spectrum in one embodiment;

[0069] Figure 14 This is a ROI region diagram of the Y-axis velocity scale in one embodiment;

[0070] Figure 15 Y-axis velocity scale diagram in one embodiment

[0071] Figure 16 Here is a digital ROI region map in one embodiment;

[0072] Figure 17 This is a schematic diagram of the split character 5 in one embodiment;

[0073] Figure 18 This is a schematic diagram illustrating the splitting character 0 in one embodiment;

[0074] Figure 19 This is a schematic diagram of the X-axis time scale in one embodiment.

[0075] Figure 20 This is a schematic diagram illustrating the visual effect of calculating the correlation coefficient of umbilical blood flow spectrum in one embodiment.

[0076] Figure 21 This is a structural block diagram of a device for measuring and parallel processing umbilical blood flow ultrasound images in one embodiment;

[0077] Figure 22 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0079] The umbilical blood flow ultrasound image measurement and parallel processing method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 is connected to ultrasound acquisition device 104, allowing medical personnel to scan the pregnant woman's abdomen and acquire fetal ultrasound images. Terminal 102 uses the umbilical blood flow ultrasound image measurement and parallel processing method of this application to process the fetal ultrasound images and obtain the umbilical blood flow spectrum correlation coefficient.

[0080] The terminal 102 may be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices.

[0081] In one embodiment, such as Figure 2 As shown, a method for umbilical blood flow ultrasound image measurement and parallel processing is provided, which can be applied to... Figure 1 Taking terminal 102 as an example, the explanation includes the following steps:

[0082] Step 202: Accelerate the acquisition of standard cross-sections of umbilical blood flow through GPU parallel computing.

[0083] Specifically, the terminal 102 identifies the fetal ultrasound image and accelerates the acquisition of the standard section of umbilical blood flow through parallel computing using the GPU.

[0084] To facilitate accurate identification of the standard section of umbilical blood flow from fetal ultrasound images, in one embodiment, an umbilical blood flow ultrasound image recognition model can be pre-trained. The standard section of umbilical blood flow in one embodiment is as follows: Figure 3 As shown.

[0085] The umbilical cord blood flow ultrasound image recognition task is defined as a classification task. A deep learning model is used to automatically identify the standard cross-section of umbilical cord blood flow. Three categories are designed: standard cross-section of umbilical cord blood flow, non-standard cross-section of umbilical cord blood flow, and other cross-sections. The Swin Transformer model is selected as the classification model, and GPU parallel computing is used to accelerate the model training and inference process. The specific strategy is as follows:

[0086] 1. Data Parallelism. The dataset is divided into samples or batches to obtain multiple subsets, each called a sub-batch. The entire deep learning model is then copied to multiple GPU units, ensuring each GPU unit has a complete copy of the inference model. The sub-batches are distributed across the GPUs for parallel processing. Each GPU unit processes one or more sub-batches, independently performing forward propagation, backpropagation, and parameter updates. Data parallelism fully utilizes the computing power of multiple GPUs, distributing tasks across multiple units for parallel processing. This improves the system's scalability and extensibility, accelerating the training and inference processes of deep learning models.

[0087] 2. Convolution Acceleration Algorithm. In traditional convolution calculations, each weight of each convolution kernel needs to be multiplied with each pixel of the input data, and then the results are added together to generate the final result. The complexity of this calculation method increases quadratically or cubically with the size of the convolution kernel and the size of the input data. The Winograd fast convolution algorithm reduces this computational complexity through a series of transformations. In the transformation phase of the Winograd algorithm, the input data and the convolution kernel are transformed, so that the original convolution operation can be transformed into a simpler matrix multiplication. The transformation of the input data mainly involves dividing the input data into non-overlapping sub-blocks and transforming it into a new representation through a series of linear transformations. The convolution kernel is also transformed into a new representation, so that the convolution operation can be completed by matrix multiplication with the transformed input data. In the multiplication phase of the Winograd algorithm, matrix multiplication is performed. By performing matrix multiplication on the transformed input data and the convolution kernel, an intermediate result can be obtained. In the inverse transformation phase of the Winograd algorithm, the intermediate result obtained by multiplication is transformed back into the original representation to obtain the final convolution output. The Winograd algorithm significantly reduces the number of multiplications and memory accesses by converting convolution operations into matrix multiplications, thereby improving computational efficiency and accelerating the training and inference process of deep learning models.

[0088] Step 204: Locate the region of interest for umbilical blood flow in the standard cross-section of the umbilical blood flow.

[0089] The region of interest for umbilical blood flow is the region that includes umbilical blood flow. Specifically, image-based methods can be used to locate the region of interest for umbilical blood flow in the standard cross-section of the umbilical blood flow.

[0090] In one embodiment, the method for locating the region of interest (ROI) of the umbilical blood flow in the standard cross-section includes: converting the input umbilical blood flow standard cross-section (RGB image) into a grayscale image; calculating a grayscale threshold based on the grayscale histogram features of the grayscale image; and performing a coarse-grained global binarization process to obtain a binarized image, such as... Figure 4 As shown. Based on the pixel features of the binarized image, the process is traversed row by row to find the dividing line between the region of interest (ROI) and irrelevant regions of umbilical blood flow, as shown. Figure 5 As shown, the ROI region of umbilical blood flow is located and segmented, as follows. Figure 6 As shown. Considering time complexity, the input RGB image is scaled when locating the umbilical blood flow ROI region to save time.

[0091] Step 206: Determine the X-axis region from the region of interest of the umbilical blood flow.

[0092] Specifically, edge features are extracted for the umbilical blood flow ROI region to find possible X-axis regions.

[0093] In one embodiment, identifying the region of interest (ROI) of the umbilical blood flow and determining the X-axis region includes: traversing the pixels of the ROI of the umbilical blood flow; if the pixel value of the current pixel is greater than the pixel value of its upper and lower adjacent pixels, then the current pixel is determined to be a point in the X-axis region; processing the pixel values ​​of the points in the X-axis region to obtain an edge feature map of the X-axis; and calculating the mean and variance of each row of pixels in the edge feature map to determine the X-axis region.

[0094] Specifically, through extensive cleaning and observation of umbilical blood flow ultrasound data, a significant difference in pixel values ​​was found between the x-axis region and the surrounding spectral regions. To address this phenomenon, the umbilical blood flow ROI region was traversed line by line, and the following method was used to distinguish and locate the x-axis position: if the pixel value of the current point differs from that of its adjacent points above and below by more than a certain threshold, the current point was selected as a potential x-axis region point, and its pixel value was set to 255; otherwise, it was set to 0. This process yielded the edge feature map, as shown below. Figure 7 As shown, the x-axis position is finally located by calculating factors such as the mean and variance of pixels in each row of the edge feature map. Figure 8 As shown.

[0095] Step 208: Determine the umbilical blood flow spectral envelope based on the region of interest of the umbilical blood flow and the X-axis region.

[0096] Among them, the umbilical blood flow spectrum envelope refers to the envelope of the umbilical blood flow spectrum curve.

[0097] In one embodiment, determining the umbilical blood flow spectral envelope based on the region of interest (ROI) and the X-axis region includes: performing a local binarization operation on the ROI to obtain a binarized image; obtaining the umbilical blood flow spectrum based on the edge information of the contour in the binarized image; dividing the umbilical blood flow spectrum into an upper region and a lower region on the binarized image, with the X-axis region as the boundary; and determining the umbilical blood flow spectral envelope from the upper and lower regions based on preset factors, including area, mean pixel value, and variance.

[0098] Specifically, for the umbilical blood flow ROI region, a "fine-grained" local binarization operation is performed based on the pixel gradient change information of the image to obtain a binarized umbilical blood flow ROI map, such as... Figure 9As shown in the figure. Local binarization is used primarily to more accurately obtain the edge information of the umbilical blood flow spectral envelope, improving the accuracy of spectral measurement. The umbilical blood flow spectrum is divided into upper and lower regions along the X-axis. Only regions exhibiting significant vertical fluctuations accompanied by stable peaks and troughs are considered valid regions. In this figure, the lower region along the X-axis is the valid region. Possible spectral regions are determined by finding the contours with the largest areas above and below the X-axis. The spectral regions are then finally located using factors such as the mean, variance, and area of ​​the pixel values. Since the calculated spectral envelope contains many redundant points, a further pruning and smoothing operation is required. For efficiency and real-time considerations, the sliding window size is 3. Finally, the valid umbilical blood flow spectral envelope is calculated, as shown in the figure. Figure 10 As shown.

[0099] Step 210: Calculate the peak and trough points of the umbilical blood flow spectrum based on the envelope of the umbilical blood flow spectrum.

[0100] In one embodiment, calculating the peaks and troughs of the umbilical blood flow spectrum based on the umbilical blood flow spectrum envelope includes: traversing the umbilical blood flow spectrum envelope to determine its highest and lowest points; shifting the highest point of the umbilical blood flow spectrum envelope to obtain a peak reference line, and shifting the lowest point to obtain a trough reference line; finding local highest points in the peak search area corresponding to the peak reference lines to obtain the peaks of the umbilical blood flow spectrum; and finding local lowest points in the trough search area corresponding to the trough reference lines to obtain the troughs of the umbilical blood flow spectrum.

[0101] Specifically, the umbilical blood flow envelope is traversed horizontally in sequence to find the highest and lowest points of the envelope. Based on this, an appropriate offset is selected, and the highest point is offset upwards to highThresh. Figure 11 (as shown by the green line), shifting downwards from the lowest point to lowThresh ( Figure 11 The red line shown (if the effective region of the spectrum is above the X-axis, the offset direction is opposite), as shown Figure 11 Find the highest local point in the region below the highThresh of the umbilical cord blood flow envelope as the peak point, and find the lowest local point in the region above the lowThresh of the umbilical cord blood flow envelope as the trough point. Record the coordinates of the peak and trough points. Figure 12 As shown.

[0102] Step 212: Select a continuous and stable umbilical blood flow spectrum based on the peaks and troughs of the umbilical blood flow spectrum.

[0103] Specifically, in ultrasound examinations, 2-3 continuous and stable spectral segments are generally selected to calculate the umbilical blood flow correlation coefficient. Stability here primarily refers to the velocity difference and time difference between adjacent peaks and troughs being less than a given threshold. Selecting continuous and stable umbilical blood flow spectra for spectral coefficient calculation effectively avoids situations where differences in flow velocity or heart rate duration within a single spectrum lead to measurements significantly deviating from the true value.

[0104] In one embodiment, selecting a continuous and stable umbilical blood flow spectrum based on the peak and trough points of the umbilical blood flow spectrum includes: treating adjacent peaks and troughs in the umbilical blood flow spectrum as a spectral unit; sequentially traversing each spectrum, and merging adjacent spectra when the velocity difference and time difference between peaks and troughs of adjacent spectra are both less than the corresponding given thresholds; placing the merged spectrum into a continuous and stable spectrum candidate queue; and selecting the spectrum with the largest peak velocity from the continuous and stable spectrum candidate queue to obtain a continuous and stable umbilical blood flow spectrum.

[0105] Specifically, the method for selecting continuous and stable umbilical blood flow spectra treats adjacent peaks and troughs as a spectral unit, iterates through each spectrum sequentially, and merges adjacent spectra that meet the conditions when the velocity difference or time difference between peaks and troughs is less than a given threshold. This spectrum is then added to a candidate queue of continuous and stable spectra. Finally, the spectrum with the highest peak velocity is selected from the candidate queue to calculate the umbilical blood flow correlation coefficient. Figure 13 As shown.

[0106] Step 214: Locate the first region of interest (ROI) of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first ROI.

[0107] Specifically, the unit in the Y-axis direction of the umbilical blood flow spectrum diagram represents the velocity unit, namely cm / s. The velocity calculation of the umbilical blood flow spectrum coefficients PSV and EDV requires knowing the conversion ratio between each pixel in the Y-axis direction and the actual flow velocity.

[0108] Specifically, the method of locating the first region of interest (ROI) of the velocity scale in the Y-axis direction and calculating the conversion ratio of the velocity scale in the Y-axis direction based on the first ROI includes: locating the first ROI of the velocity scale in the Y-axis direction; locally binarizing the first ROI of the velocity scale to obtain a binary image of the Y-axis velocity scale; determining candidate Y-axis velocity scale points by finding connected components in the binary image of the Y-axis velocity scale; filtering noise scale points among the candidate Y-axis velocity scale points by length, width, and the interval distance between connected components to obtain Y-axis velocity scale points; sequentially traversing the Y-axis velocity scale points, and if there are numbers near the Y-axis velocity scale point, determining the velocity scale point as a digital region of interest; performing local binarization on the digital region of interest, segmenting each digit character using pixel gradient change information, and identifying the actual flow velocity represented by the number; and obtaining the conversion ratio of the velocity scale in the Y-axis direction based on the actual flow velocity and the position coordinates of the Y-axis.

[0109] Specifically, to calculate the conversion ratio, first locate the first ROI region of the Y-axis velocity scale, such as... Figure 14 As shown, a method similar to that used for locating the ROI region of umbilical blood flow is employed. By locating the position of the Y-axis velocity scale, the location of the numerical ROI region (where the numbers appear near the velocity scale points) can be quickly determined. By identifying the numbers in the image, the conversion ratio between each pixel in the image and the actual flow velocity can be determined. The specific method is as follows: First, the velocity scale ROI region is locally binarized. Velocity scale points are located by finding connected components in the image. Noisy scale points are filtered out using factors such as length, width, and the distance between connected components. Finally, the velocity scale points are calculated. Figure 15 As shown. Next, iterate through the velocity scale points from top to bottom, checking for numbers near each velocity scale point. If a number is found, locate the ROI region containing that number, such as... Figure 16 As shown. Within the digit ROI region, a fine-grained local binarization operation is performed again, utilizing pixel gradient change information to segment each digit character, as shown. Figure 17 The separator character 5 shown, and as shown Figure 18 The delimiter '0' shown is used to identify numbers using template matching for each digit. The identified number represents the actual flow velocity at the corresponding scale point. Combined with the previously calculated X-axis position coordinates, the conversion ratio between each pixel in the image and the actual flow velocity can be calculated.

[0110] Step 216: Locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the region of interest.

[0111] In the umbilical blood flow spectrum diagram, the unit in the X-axis direction represents the time unit, i.e., s. The calculation of the umbilical blood flow spectrum coefficients RI and HR requires knowing the conversion ratio of each pixel in the X-axis direction to the actual time.

[0112] Specifically, locating the second region of interest (ROI) of the time scale in the X-axis direction and calculating the conversion ratio of the time scale in the X-axis direction based on the ROI includes: locating the second ROI of the time scale in the X-axis direction; locally binarizing the second ROI of the time scale to obtain a binary image of the X-axis time scale; determining candidate X-axis time scale points by finding connected components in the binary image of the X-axis time scale; filtering out noise scale points among the candidate X-axis time scale points by using length, width, and the interval distance between connected components to obtain X-axis time scale points; and calculating the interval pixels between adjacent X-axis time scale points to obtain the conversion ratio of the time scale in the X-axis direction.

[0113] Specifically, the method for calculating the conversion ratio is similar to the method for calculating the conversion ratio of the velocity scale in the Y-axis direction. The final result is a time scale diagram, as shown below. Figure 19 As shown, there is no need to identify numbers here because the time interval between adjacent scale points is fixed. Taking the umbilical cord blood flow spectrum image of the Samsung model as an example, the time interval between each adjacent scale point is 200ms. Therefore, we only need to calculate the number of pixels between adjacent scale points to calculate the conversion ratio between pixels in the image and actual time.

[0114] Step 218: Based on the continuous and stable umbilical blood flow spectrum, the conversion ratio of the velocity scale in the Y-axis direction, and the conversion ratio of the time scale in the X-axis direction, obtain the correlation coefficient of the umbilical blood flow spectrum.

[0115] Specifically, the correlation coefficients of umbilical blood flow spectrum include PSV (peak systolic velocity), EDV (end diastolic velocity), S / D coefficient of systolic / diastolic ratio, resistance index (RI), pulsatility index (PI), and HR coefficient (umbilical blood flow fetal heart rate).

[0116] Specifically, the PSV is calculated as follows: select the point with the highest peak in the continuous and stable umbilical blood flow spectrum as the PSV calculation point, and multiply the relative coordinate of the point in the Y-axis direction by the conversion ratio of the velocity scale in the Y-axis direction to obtain the PSV velocity.

[0117] The EDV is calculated as follows: the trough immediately to the right of the peak corresponding to the PSV is used as the calculation point. The calculation method for EDV velocity is the same as that for PSV velocity. Specifically, the trough immediately to the right of the peak corresponding to the PSV is used as the EDV calculation point. The EDV is obtained by multiplying the relative coordinate of the EDV calculation point in the Y-axis direction by the conversion ratio of the velocity scale in the Y-axis direction.

[0118] The formula for calculating the systolic-to-diastolic ratio (S / D coefficient) is PSV / EDV.

[0119] The formula for calculating the resistance index RI is: (PSV - EDV) / PSV.

[0120] The formula for calculating the pulsatility index (PI) is (PSV - EDV) / ,in The average velocity represents the continuous and stable umbilical blood flow spectrum, and its calculation method is similar to that of PSV.

[0121] The heart rate (HR) coefficient represents the number of heartbeats per minute. A spectrum containing only one peak and trough represents one heartbeat. The HR coefficient is calculated by subtracting the X-axis coordinate of the starting point from the X-axis coordinate of the ending point of the continuous stable umbilical blood flow spectrum, multiplied by the conversion ratio of the time scale in the X-axis direction. This gives the total time taken for the continuous stable umbilical blood flow spectrum. Then, the total number of heartbeats during the continuous stable umbilical blood flow spectrum is counted to calculate the time taken for each heartbeat. Finally, this is converted into heartbeats per minute to calculate the HR. The final visualization of the umbilical blood flow spectrum correlation coefficient is shown below. Figure 20 As shown.

[0122] In this embodiment, based on the standard cross-section of umbilical blood flow identified by the deep learning model Swin Transformer, digital image processing technology can be used to measure the relevant growth parameters of the umbilical blood flow spectrum in real time.

[0123] The umbilical blood flow ultrasound image measurement and parallel processing method of this application utilizes digital image processing technology to identify the region of interest (ROI), X-axis region, and umbilical blood flow spectral envelope of the standard umbilical blood flow section. This allows for the calculation of peaks and troughs in the umbilical blood flow spectrum, and the localization of a continuous and stable umbilical blood flow spectrum. Further identification of scale points enables rapid localization of velocity and time scales, thereby efficiently calculating the conversion ratios of the Y-axis velocity scale and the X-axis time scale. Based on this, using the calculated continuous and stable umbilical blood flow spectrum, the conversion ratios of the Y-axis velocity scale, and the conversion ratios of the X-axis time scale, the correlation coefficient of the umbilical blood flow spectrum can be automatically and accurately measured. This process requires no manual intervention, significantly improving the calculation efficiency of the umbilical blood flow spectrum correlation coefficient.

[0124] This method, based on digital image processing technology, enables automated measurement of umbilical cord blood flow spectral coefficients, achieving real-time measurement with a processing speed of up to 90 FPS. It utilizes rapid positioning and identification of scale points to efficiently and accurately calculate the conversion ratio between each pixel and flow velocity. This method automates umbilical cord blood flow spectral coefficient measurement, effectively improving measurement precision and accuracy.

[0125] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0126] Based on the same inventive concept, this application also provides an umbilical blood flow ultrasound image measurement and parallel processing apparatus for implementing the aforementioned umbilical blood flow ultrasound image measurement and parallel processing method. The solution provided by this apparatus is similar to the solution described in the above method; therefore, the specific limitations of one or more embodiments of the umbilical blood flow ultrasound image measurement and parallel processing apparatus provided below can be found in the limitations of the umbilical blood flow ultrasound image measurement and parallel processing method described above, and will not be repeated here.

[0127] In one embodiment, such as Figure 21 As shown, a device for measuring and parallel processing umbilical blood flow ultrasound images is provided, comprising:

[0128] The section acquisition module 2101 is used to accelerate the acquisition of standard sections of umbilical blood flow through GPU parallel computing;

[0129] The region of interest (ROI) localization module 2102 is used to locate the umbilical blood flow region of interest in the standard cross-section of the umbilical blood flow.

[0130] X-axis analysis module 2103 is used to determine the X-axis region from the region of interest of the umbilical blood flow;

[0131] Envelope analysis module 2104 is used to determine the umbilical blood flow spectrum envelope based on the region of interest of the umbilical blood flow and the X-axis region;

[0132] The first calculation module 2105 is used to calculate the peak and trough points of the umbilical blood flow spectrum based on the envelope of the umbilical blood flow spectrum.

[0133] The spectrum selection module 2106 is used to select a continuous and stable umbilical blood flow spectrum based on the peak and trough points of the umbilical blood flow spectrum.

[0134] The first conversion ratio calculation module 2107 is used to locate the first region of interest of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first region of interest.

[0135] The second conversion ratio calculation module 2108 is used to locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the region of interest.

[0136] The coefficient calculation module 2109 is used to obtain the correlation coefficient of the umbilical blood flow spectrum based on the continuous and stable umbilical blood flow spectrum, the conversion ratio of the velocity scale in the Y-axis direction, and the conversion ratio of the time scale in the X-axis direction.

[0137] Each module in the aforementioned umbilical blood flow ultrasound image measurement and parallel processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0138] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 22 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a method for umbilical cord blood flow ultrasound image measurement and parallel processing. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device casing, or an external keyboard, touchpad, or mouse.

[0139] Those skilled in the art will understand that Figure 22 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0140] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the umbilical blood flow ultrasound image measurement and parallel processing method of the above embodiments.

[0141] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the umbilical blood flow ultrasound image measurement and parallel processing method of the above embodiments.

[0142] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the umbilical blood flow ultrasound image measurement and parallel processing method of the above embodiments.

[0143] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0144] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0145] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for measuring and parallel processing umbilical blood flow ultrasound images, characterized in that, The method includes: Accelerate the acquisition of standard cross-sections of umbilical blood flow through GPU parallel computing; Locate the region of interest for umbilical blood flow in the standard section of the umbilical blood flow; Determine the X-axis region from the region of interest of the umbilical blood flow; The umbilical blood flow spectral envelope is determined based on the region of interest of the umbilical blood flow and the X-axis region; Calculate the peak and trough points of the umbilical blood flow spectrum based on the umbilical blood flow spectrum envelope; Based on the peaks and troughs of the umbilical blood flow spectrum, a continuous and stable umbilical blood flow spectrum is selected. Locate the first region of interest (ROI) of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first ROI. Locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the second region of interest; Based on the continuous and stable umbilical blood flow spectrum, the conversion ratio of the velocity scale in the Y-axis direction, and the conversion ratio of the time scale in the X-axis direction, the correlation coefficient of the umbilical blood flow spectrum is obtained. The first region of interest for locating the velocity scale in the Y-axis direction, and calculating the conversion ratio of the velocity scale in the Y-axis direction based on the first region of interest, includes: Locate the first region of interest for the velocity scale along the Y-axis; The first region of interest of the velocity scale is locally binarized to obtain a binary image of the Y-axis velocity scale. Candidate Y-axis velocity scale points are determined by finding connected components in the binary image of the Y-axis velocity scale. The Y-axis velocity scale points are obtained by filtering out the noise scale points in the candidate Y-axis velocity scale points by the length, width and the interval distance between connected components. The Y-axis velocity scale points are traversed sequentially. If there are numbers near the Y-axis velocity scale point, then the velocity scale point is determined to be a region of interest for numbers. Local binarization is performed on the digital region of interest. Pixel gradient change information is used to segment each digital character and identify the actual flow rate represented by the number. Based on the actual flow velocity and the position coordinates of the Y-axis, the conversion ratio of the velocity scale in the Y-axis direction is obtained; The second region of interest for locating the time scale in the X-axis direction, and calculating the conversion ratio of the time scale in the X-axis direction based on the second region of interest, includes: Locate the second region of interest on the time scale along the X-axis; The second region of interest of the time scale is locally binarized to obtain a binary image of the X-axis time scale. Candidate X-axis time scale points are determined by searching for connected components in the binary image of the X-axis time scale. The X-axis time scale points are obtained by filtering out the noise scale points in the candidate X-axis time scale points by length, width, and interval distance between connected components. Calculate the pixel interval between adjacent X-axis time scale points, and combine the time interval between adjacent interval scale points to obtain the conversion ratio of the X-axis time scale. The umbilical blood flow spectrum correlation coefficients include PSV, EDV, the S / D coefficient of the systolic-diastolic ratio, the resistance index, the pulsatility index, and HR; The correlation coefficient of the umbilical blood flow spectrum is obtained based on the conversion ratio of the continuous and stable umbilical blood flow spectrum, the velocity scale in the Y-axis direction, and the time scale in the X-axis direction, including: The point with the highest peak in the continuous and stable umbilical blood flow spectrum is selected as the PSV calculation point. The PSV is obtained by multiplying the relative coordinate of the PSV calculation point in the Y-axis direction by the conversion ratio of the velocity scale in the Y-axis direction. The trough point immediately to the right of the peak point corresponding to PSV is used as the EDV calculation point. The EDV is obtained by multiplying the relative coordinate of the EDV calculation point in the Y-axis direction by the conversion ratio of the velocity scale in the Y-axis direction. The formula for calculating the systolic-to-diastolic ratio (S / D coefficient) is PSV / EDV; The formula for calculating the resistance index is: (PSV - EDV) / PSV; The formula for calculating the pulsatility index is (PSV - EDV) / ,in The average velocity representing a continuous and stable spectrum of umbilical blood flow; The HR is calculated as follows: subtract the X-axis coordinate of the starting point from the X-axis coordinate of the end point of the continuous and stable umbilical blood flow spectrum, multiplied by the conversion ratio of the time scale in the X-axis direction, to obtain the total time used for the continuous and stable umbilical blood flow spectrum. Then, count the number of heartbeats in the continuous and stable umbilical blood flow spectrum to obtain the time used for each heartbeat. Convert the time used for each heartbeat to the number of heartbeats per minute to obtain the HR.

2. The method according to claim 1, characterized in that, Determining the X-axis region from the region of interest of the umbilical blood flow includes: The pixels in the region of interest of the umbilical blood flow are traversed. If the pixel value of the current pixel is greater than the pixel value of the pixels in the upper and lower adjacent positions, the current pixel is determined to be a point in the X-axis region. The pixel values ​​of the points in the X-axis region are processed to obtain the edge feature map of the X-axis; Calculate the mean and variance of each row of pixels in the edge feature map to determine the X-axis region.

3. The method according to claim 1, characterized in that, The step of determining the umbilical blood flow spectral envelope based on the region of interest of the umbilical blood flow and the X-axis region includes: The region of interest in the umbilical blood flow is locally binarized to obtain a binarized image; The umbilical blood flow spectrum is obtained based on the edge information of the contour in the binarized image; The binarized image is divided into an upper region and a lower region, with the X-axis region as the boundary; The umbilical blood flow spectrum envelope is determined from the upper and lower regions based on preset factors; the preset factors include area, mean pixel value, and variance.

4. The method according to claim 1, characterized in that, The step of calculating the peaks and troughs of the umbilical blood flow spectrum based on the umbilical blood flow spectrum envelope includes: Traverse the umbilical blood flow spectrum envelope to determine the highest and lowest points of the umbilical blood flow spectrum envelope; Based on the highest and lowest points of the umbilical blood flow spectrum envelope, the highest point of the umbilical blood flow spectrum envelope is offset to obtain a peak reference line, and the lowest point of the umbilical blood flow spectrum envelope is offset to obtain a trough reference line. In the peak search area corresponding to the peak reference line, find the local highest point to obtain the peak point of the umbilical blood flow spectrum; Within the valley search area corresponding to the valley reference line, find the local lowest point to obtain the valley point of the umbilical blood flow spectrum.

5. The method according to claim 1, characterized in that, The step of selecting a continuous and stable umbilical blood flow spectrum based on the peaks and troughs of the umbilical blood flow spectrum includes: Adjacent peaks and troughs in the umbilical blood flow spectrum are considered as a single spectral unit. The spectrum is traversed sequentially, and when the velocity difference and time difference between the peaks and troughs of adjacent spectra are both less than the corresponding given threshold, the adjacent spectra are merged. The merged spectrum is added to the continuous stable spectrum candidate queue; The spectrum with the highest peak velocity is selected from the continuous stable spectrum candidate queue to obtain the continuous stable umbilical blood flow spectrum.

6. A device for measuring and parallel processing umbilical blood flow ultrasound images, used to implement the method for measuring and parallel processing umbilical blood flow ultrasound images according to any one of claims 1 to 5, characterized in that, The device includes: The section acquisition module is used to accelerate the acquisition of standard sections of umbilical blood flow through GPU parallel computing; The region of interest (ROI) localization module is used to locate the ROI of the umbilical blood flow in the standard cross-section of the umbilical blood flow. The X-axis analysis module is used to determine the X-axis region from the region of interest of the umbilical blood flow. The envelope analysis module is used to determine the umbilical blood flow spectrum envelope based on the region of interest of the umbilical blood flow and the X-axis region; The first calculation module is used to calculate the peak and trough points of the umbilical blood flow spectrum based on the envelope of the umbilical blood flow spectrum. The spectrum selection module is used to select a continuous and stable umbilical blood flow spectrum based on the peak and trough points of the umbilical blood flow spectrum. The first conversion ratio calculation module is used to locate the first region of interest of the velocity scale in the Y-axis direction, and calculate the conversion ratio of the velocity scale in the Y-axis direction based on the first region of interest. The second conversion ratio calculation module is used to locate the second region of interest of the time scale in the X-axis direction, and calculate the conversion ratio of the time scale in the X-axis direction based on the second region of interest. The coefficient calculation module is used to obtain the correlation coefficient of the umbilical blood flow spectrum based on the continuous and stable umbilical blood flow spectrum, the conversion ratio of the velocity scale in the Y-axis direction, and the conversion ratio of the time scale in the X-axis direction.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

Images