Automatic real-time envelope line drawing method in PW mode of ultrasonic equipment and ultrasonic equipment

Abstract

The invention relates to the technical field of medical imaging and signal processing, in particular to an automatic real-time envelope line drawing method of ultrasonic equipment in a PW mode and the ultrasonic equipment. RF data are collected through an ultrasonic probe, and log compression and baseline translation processing are carried out to form gray scale data; calculating mathematical expectation and variance of noise through a noise evaluation module; a boundary calculation module obtains a boundary for each frame of image data to separate an upper envelope line and a lower envelope line; the signal extraction module extracts a signal main body part and removes noise; the envelope line calculation module determines edge points of an envelope line through accumulation and calculation, and a continuous and smooth envelope line is generated through quadratic curve fitting and a least square method. According to the scheme, the envelope line drawing precision and stability are effectively improved, noise interference is removed, efficient and real-time envelope line drawing is achieved, and reliable technical support is provided for ultrasonic equipment.

Description

Technical Field

[0001] This invention relates to the fields of medical imaging and signal processing technology, and in particular to an automatic real-time envelope plotting method and ultrasound equipment in PW mode. Background Technology

[0002] In PW mode of ultrasound equipment, automated real-time envelope mapping is crucial for accurately detecting and displaying blood flow velocity. This technology helps doctors quickly and accurately assess the health of a patient's blood vessels, enabling timely diagnosis and treatment. However, current automated real-time envelope mapping technology has some limitations, primarily due to the following reasons:

[0003] Ultrasonic signals are susceptible to noise during transmission, resulting in highly noisy data that can affect the accurate plotting of the envelope. Due to the high frequency and high resolution of ultrasound data, data processing is highly complex, requiring effective methods to smooth and extract the main signal components. Real-time processing necessitates fast and efficient algorithms to ensure timely data processing and display.

[0004] Existing technologies have the following shortcomings or defects when facing these problems: Existing noise processing methods often fail to completely remove noise, resulting in less smooth and accurate envelope drawing. Many existing methods are inefficient in processing data, failing to meet the requirements of real-time processing and affecting the effectiveness of ultrasound equipment. Due to the limitations of data processing algorithms, the envelope drawn by existing methods often exhibits jagged edges and abrupt changes, affecting the accurate assessment of blood flow velocity. Summary of the Invention

[0005] To address the numerous problems existing in the prior art, this invention provides an automatic real-time envelope drawing method and an ultrasound device in PW mode. The method involves performing log compression and baseline shifting on the RF data acquired by the ultrasound device to form grayscale data. Then, a noise assessment module calculates the mathematical expectation and variance of the noise, and determines the boundary line for each frame of image data to separate the upper and lower envelopes. A signal extraction module extracts the main signal from the image data and removes noise interference. The envelope calculation module further accumulates and calculates the data in each column to determine the edge points of the envelope, and generates a continuous and smooth envelope through quadratic curve fitting and the least squares method.

[0006] An automatic real-time envelope plotting method for ultrasound equipment in PW mode includes the following steps:

[0007] In PW mode, an ultrasound probe emits ultrasound waves towards the target blood vessel tissue and continuously acquires RF data; in ultrasound PW (pulse wave) mode, the ultrasound probe emits ultrasound waves towards the target blood vessel tissue. When ultrasound waves encounter blood flow within the blood vessel, the Doppler effect occurs, and the reflected signal contains information about the blood flow velocity.

[0008] The acquired RF data is log-compressed to form grayscale data, and the expected value and variance of the noise are calculated based on this grayscale data. These reflected signals are received by the probe and converted into RF (radio frequency) data. Due to the massive amount of data in PW mode, effective processing and compression of this data are necessary. In actual acquisition, various noises are mixed into the data, affecting image quality and subsequent processing steps. Therefore, noise modeling and evaluation are required. Assuming the noise follows a Gaussian distribution, the characteristics of the noise can be effectively described by calculating its expected value and variance.

[0009] For each frame of image data, a boundary line is determined to separate the upper and lower envelopes. This boundary line divides the image into upper and lower parts, facilitating subsequent envelope extraction. A weight vector is generated by calculating local averages and binarizing it. Then, a convolution is performed between the binarized vector and the weight vector; the maximum value of the convolution is used as the boundary line.

[0010] The main signal component is extracted from each frame of image data, and the expected value of the noise is subtracted to obtain the processed image data. The main signal component is extracted from each frame of image data, and the influence of noise is removed. By subtracting the expected value of the noise, cleaner image data can be obtained.

[0011] For each frame of image data, calculate the edge points of the upper and lower envelopes for each column of data. The upper and lower envelopes represent the maximum and minimum values ​​of blood flow velocity, respectively.

[0012] The obtained edge points are denoised using ordered point sets, and a quadratic curve fitting method based on local neighborhoods is employed to generate a continuous and smooth envelope. The edge points of the obtained envelope are then denoised. The quadratic curve fitting method generates a continuous and smooth envelope, eliminating jagged edges and abrupt changes.

[0013] This invention, through a series of sophisticated algorithms and processing steps, enables the automatic and real-time mapping of blood flow velocity envelopes in ultrasound PW mode. These steps include data acquisition and preprocessing, noise assessment, boundary line calculation, signal core extraction, envelope edge point calculation, and ordered point set denoising. Through these steps, the generated envelope accurately and stably reflects the boundary information of blood flow velocity, providing crucial support for medical diagnosis.

[0014] Preferably, the grayscale data is subjected to baseline shifting so that the baseline is located in the center of the image.

[0015] Baseline shifting is a signal processing technique used to adjust the baseline position of image data. In ultrasound PW mode, RF data is compressed using log compression to form a grayscale data matrix. The vertical data in the matrix represents frequency or blood flow velocity information, while the horizontal data represents time information. During actual acquisition, due to various noises and system offsets, the baseline of the grayscale data (i.e., the baseline for frequency information) may not be centered in the image. This causes an overall image shift, affecting the accuracy of subsequent processing steps.

[0016] Baseline shifting eliminates errors caused by system offset by calculating the baseline position of the grayscale data and adjusting it to the center of the image. The specific steps are as follows:

[0017] The baseline position of the current grayscale data is calculated using statistical analysis methods. This can be achieved by calculating the mean or median of the grayscale data. Based on the calculated baseline position, the grayscale data is shifted as a whole to align the baseline with the center of the image.

[0018] This invention eliminates data fluctuations caused by system offset by adjusting the baseline to the center of the image, resulting in more stable data. This is crucial for steps such as noise assessment and signal extraction. Baseline shifting makes the overall image structure more symmetrical, facilitating image analysis and processing by doctors and algorithms, and improving image readability and accuracy. Due to baseline alignment, the correspondence between frequency and time information becomes clearer, making the calculation of the upper and lower envelopes more accurate. The final generated envelope more accurately reflects the true blood flow velocity.

[0019] Preferably, the calculation steps for the boundary line include:

[0020] Calculate local averages; calculate the average value for a local region of each frame of image data. Local averages can effectively smooth image data, reduce the influence of local noise, and thus highlight the main features of the signal. By using a sliding window method, the average value is calculated for each small region of the image data, resulting in a new average matrix.

[0021] Binarize the local average values; perform binarization on the calculated local average value matrix. Binarization converts continuous grayscale values ​​into discrete values ​​of 0 and 1 to facilitate subsequent convolution calculations. Select a threshold, set the local average values ​​above the threshold to 1, and those below the threshold to 0, forming a binarized matrix.

[0022] Generate a weight vector; generate a weight vector based on the binarization matrix. The weight vector is used to highlight the signal components relative to the noise components, enhancing the signal characteristics. The binarization matrix is ​​processed to generate a vector containing the weights; the length of the weight vector is the same as the number of columns in the binarization matrix.

[0023] Perform convolution on the binarized vector and the weight vector, and use the maximum value of the convolution as the boundary. The boundary is determined by selecting the maximum point in the convolution result vector. The maximum point corresponds to the part of the signal with the most obvious features and can effectively separate the upper and lower envelopes. Find the maximum point in the convolution result vector; this point is the desired boundary.

[0024] By calculating local averages and smoothing noise, signal characteristics are enhanced, and the interference of noise on envelope calculation is reduced. Binarization and weight vector generation simplify the data processing, facilitating subsequent convolution operations. Finding the maximum point as the dividing line through convolution operations effectively separates the upper and lower envelopes, making envelope calculation clearer and more accurate.

[0025] Preferably, the step of extracting the main body of the signal includes:

[0026] Averaging the image data over time is crucial because random noise exists in the acquired image data. Averaging the image data over time effectively smooths the data, reduces the impact of random noise, and enhances signal stability. The values ​​at each time point of each frame of image data are averaged to obtain a time-averaged data matrix. This step smooths the data by calculating the mean of each pixel on the time axis, reducing noise interference in subsequent processing.

[0027] Subtracting the expected value of the noise yields the processed image data. The expected value of the noise represents its average value. By subtracting this expected value, the influence of noise can be further eliminated, and the main part of the signal can be extracted. This processed data better reflects the true blood flow velocity information. Subtracting the previously calculated expected value of the noise from the time-averaged data matrix yields the noise-reduced processed data. This step, by eliminating the average noise value, makes the processed data clearer and more accurate.

[0028] Preferably, the step of calculating the edge points of the envelope includes:

[0029] The upper and lower envelope data are extracted based on the dividing line; the dividing line divides the image data into an upper and lower half, corresponding to the upper and lower envelopes respectively. By extracting these two parts of data, the edge points of the upper and lower envelopes can be calculated separately. Using the previously calculated dividing line, each frame of image data is divided into upper and lower parts, and the data of these two parts is extracted separately in preparation for further calculations.

[0030] Calculate the sum of the first n terms of each column of data. The sum of the first n terms of each column reflects the cumulative characteristics of the data. By calculating the cumulative sum, the data can be further smoothed and the impact of noise can be reduced. For each column of data, calculate the sum of the first n terms starting from the top to obtain a matrix of cumulative sums, which is used for subsequent score calculation.

[0031] Calculate the score values, selecting the maximum point as the outer envelope and the minimum point as the inner envelope. By calculating the score values, we can determine the points in the data that best reflect the signal boundaries. The maximum point corresponds to the outer envelope, representing the upper boundary of the signal; the minimum point corresponds to the inner envelope, representing the lower boundary of the signal. Calculate the score values ​​for the cumulative sum matrix, and select the maximum and minimum points based on the score values ​​as the edge points of the outer and inner envelopes, respectively.

[0032] Preferably, the ordered point set denoising process includes:

[0033] For any point in an ordered point set, extract 16 points from its vicinity to form a new ordered point set. An ordered point set refers to the set of points representing the edges of the envelope after processing. To remove local noise, each point needs local processing by extracting points from its vicinity to form a new ordered point set. For any point in the ordered point set, extract 16 points from its vicinity to form a new locally ordered point set. These points include the points before and after the current point, making the features of the local region more prominent.

[0034] Finding a quadratic curve to fit an ordered set of points is a common signal processing method. By finding the most suitable quadratic curve to fit the ordered set of points, noise can be effectively removed and the signal smoothed. Within the extracted locally ordered set of points, a quadratic curve is found to fit these points, ensuring that the curve smoothly passes through all points, thus reducing the impact of noise on the signal.

[0035] The least squares method is used to solve for the parameters of the quadratic curve fitting. Least squares is a mathematical optimization method that aims to minimize the sum of squared errors between the fitted curve and the data points. By using least squares, the parameters of the quadratic curve can be accurately solved, resulting in the best fitting effect. Using least squares to solve for the parameters z, b, and c of the fitted curve ensures that the fitted curve approximates the locally ordered set of points as closely as possible, thus smoothing out noise.

[0036] Generate a continuous and smooth envelope after fitting. Using the fitted quadratic curves, a continuous and smooth envelope is generated, eliminating jagged edges and abrupt changes, making the envelope more meaningful and interpretable. Connecting all the fitted quadratic curves forms a single, continuous and smooth envelope, ensuring the integrity and continuity of the signal.

[0037] This invention effectively smooths noise and enhances signal continuity and stability through quadratic curve fitting in local regions. The quadratic curve fitting makes the envelope more continuous and smooth, eliminating jagged edges and abrupt changes caused by noise, and improving the practical application value of the envelope. The continuous and smooth envelope can more accurately reflect changes in blood flow velocity, providing more reliable data support and clearer images for medical diagnosis.

[0038] An ultrasonic device, comprising:

[0039] An ultrasound probe is used to emit ultrasound waves to target blood vessels and continuously acquire RF data in PW mode. Operating in PW mode, the probe acquires RF (radio frequency) data by emitting ultrasound waves and receiving the waveform signals reflected back from the target blood vessels. PW mode (pulse wave mode) is used to detect blood flow velocity and direction. The ultrasound probe can provide high-precision blood flow velocity information, monitor vascular health in real time, and help doctors make accurate diagnoses.

[0040] The data processing module, connected to the ultrasound probe, performs log compression and baseline shifting on the acquired RF data to center the baseline in the image, forming grayscale data. The module first performs log compression on the acquired RF data to reduce its dynamic range for easier visualization and subsequent processing. Then, baseline shifting is performed to adjust the baseline to the center of the image, thus forming a standardized grayscale data matrix. Log compression improves data visualization, while baseline shifting ensures data consistency and stability, reducing the impact of system offset on the image.

[0041] A noise assessment module, connected to the data processing module, is used to calculate the expected value and variance of the noise based on the grayscale data. The noise assessment module evaluates the noise level by calculating the expected value and variance of the noise in the grayscale data. Assuming the noise follows a Gaussian distribution, the mean and standard deviation of the noise are calculated using statistical methods. Noise assessment enables the system to accurately understand the noise level in the image, providing necessary basic information for subsequent signal extraction and processing, and enhancing the robustness of the overall processing.

[0042] The boundary line calculation module, connected to the noise evaluation module, is used to calculate the boundary line for each frame of image data to separate the upper and lower envelopes. The specific steps include calculating the local average value, binarizing, generating a weight vector, and performing a convolution operation. Finally, the boundary line is determined by the maximum value of the convolution. Accurate boundary line calculation ensures a clear separation between the upper and lower envelopes, making the envelope calculation more accurate and reducing the intersection and overlap of the upper and lower envelopes.

[0043] The signal extraction module, connected to the boundary line calculation module, is used to extract the main body of the signal from each frame of image data, subtract the expected value of the noise, and obtain the processed image data. Specifically, the signal extraction module extracts the main body of the signal from each frame of image data by averaging the image data over time and subtracting the expected value of the noise, resulting in noise-removed processed data. The extracted signal body is clearer, with most noise interference removed, ensuring signal purity and providing a high-quality data foundation for envelope calculation.

[0044] The envelope calculation module, connected to the signal extraction module, calculates the edge points of the upper and lower envelopes for each column of data in each frame of image data. It then performs ordered point set denoising on the obtained edge points and generates a continuous and smooth envelope using a quadratic curve fitting method based on local neighborhood. Specifically, the envelope calculation module extracts data based on the boundary line, calculates the sum of the first n terms, calculates the score, and selects the maximum value point as the outer envelope and the minimum value point as the inner envelope. Then, it performs ordered point set denoising on the obtained edge points and generates a continuous and smooth envelope using a quadratic curve fitting method based on local neighborhood. The generated envelope is smoother and more continuous, removing noise and sharp abrupt changes, improving the practical application value and accuracy of the envelope, and better reflecting the true changes in blood flow velocity.

[0045] The ultrasound equipment in this solution utilizes the collaborative work of multiple modules. From RF data acquisition by the ultrasound probe to data processing, noise assessment, boundary line calculation, signal extraction, and envelope calculation, each step ensures data accuracy and processing precision from a fundamental perspective. The resulting envelope provides clear and accurate blood flow velocity information, offering reliable support for medical diagnosis.

[0046] Preferably, the boundary line calculation module includes:

[0047] Local average value calculation unit, used to calculate local average values;

[0048] The binarization unit is connected to the local average value calculation unit and is used to binarize the local average value.

[0049] A weight vector generation unit, connected to the binarization unit, is used to generate weight vectors;

[0050] The convolution calculation unit, connected to the weight vector generation unit, is used to perform convolution on the binary vector and the weight vector, with the maximum value of the convolution serving as the boundary line.

[0051] Local averaging is a technique used to smooth image data. By calculating the average value of a local region, the influence of random noise can be reduced, highlighting the main features of the signal. In ultrasound equipment, local averaging helps smooth image data and enhance signal stability. Specifically, the local averaging unit processes a local region of the image data, calculating the average value of each pixel within that region, thereby forming a new smoothed data matrix.

[0052] Binarization is the process of converting a grayscale image into a binary image. By setting a threshold, pixel values ​​above the threshold are set to 1, and values ​​below the threshold are set to 0. This step simplifies data processing and highlights signal features. In ultrasound equipment, the binarization unit processes local average values, converting the data into a binary image of 0s and 1s to facilitate subsequent convolution calculations.

[0053] Weight vector generation involves analyzing binarized image data to generate weight vectors for convolution calculations. The purpose of the weight vectors is to highlight the signal components and reduce the influence of noise during convolution. In ultrasound equipment, the weight vector generation unit analyzes the binarized image data and generates a matching weight vector for use in convolution calculations.

[0054] Convolution is a signal processing technique that highlights specific features of a signal by performing a convolution operation with a convolution kernel. The convolution unit performs a convolution operation on a binary vector and a weight vector, and the result is used to determine the boundary line. In ultrasound equipment, the convolution unit finds the maximum value of the convolution result by performing a convolution operation on the binary vector and the weight vector, thus determining the boundary line.

[0055] Preferably, the signal extraction module includes:

[0056] The average value calculation unit is used to average the image data over time.

[0057] The noise subtraction unit, connected to the average value calculation unit, is used to subtract the mathematical expectation of the noise to obtain the processed image data.

[0058] Averaging is a fundamental method in signal processing used to smooth data and reduce the effects of random noise. In ultrasound equipment, averaging image data over time can eliminate short-term fluctuations and noise in a single frame.

[0059] Ultrasound data acquired in PW mode is a continuous time series, containing a large number of instantaneous changes and noise. By calculating the average value at each time point, these changes can be effectively smoothed out, highlighting the main signal.

[0060] For each frame of image data, the average value of each pixel is calculated along the time direction. Specifically, a sliding window method can be used to average multiple values ​​before and after each time point.

[0061] The noise subtraction unit further removes noise components from the image data by subtracting the expected value of the noise. The expected value of the noise represents the average level of the noise; subtracting this expected value can effectively reduce the impact of the noise. The expected value of the noise has already been calculated in the noise evaluation module; by subtracting it from the image data, the denoised signal can be obtained.

[0062] By subtracting the mathematical expectation of the noise, the noise component in the image data is further reduced, making the signal cleaner. After denoising, the accuracy and reliability of the signal are improved, providing a high-quality data foundation for subsequent envelope calculations.

[0063] Preferably, the envelope calculation module includes:

[0064] The data extraction unit is used to extract the upper envelope and lower envelope portions of the data based on the boundary line;

[0065] The data extraction unit divides the image data into upper and lower envelope portions based on the dividing line. The dividing line separates the two different regions so that their envelopes can be calculated separately. In ultrasound equipment, after the dividing line segments the image data, the data extraction unit extracts the data from the upper portion (corresponding to the upper envelope) and the lower portion (corresponding to the lower envelope).

[0066] The cumulative sum calculation unit, connected to the data extraction unit, is used to calculate the sum of the first n terms of each column of data. The cumulative sum calculation smooths the data, reduces the impact of random fluctuations and noise, and enhances signal stability. In ultrasound equipment, the cumulative sum calculation unit processes the extracted envelope portion of the data and calculates the cumulative sum of each column of data to facilitate subsequent score calculation.

[0067] A score calculation unit, connected to the cumulative sum calculation unit, is used to calculate the score value, selecting the maximum value point as the outer envelope and the minimum value point as the inner envelope. The score calculation unit calculates the score value for each column of data, selecting the maximum and minimum score points as the edge points of the outer and inner envelopes, respectively. The score value reflects the signal strength and characteristics. In ultrasound equipment, the score calculation unit processes the cumulative sum data to determine the edge points of the envelope.

[0068] The noise reduction unit is used to extract 16 nearby points from any point in the ordered point set to form an ordered point set, find a quadratic curve to fit the ordered point set, and use the least squares method to solve the fitting parameters of the quadratic curve to generate a continuous and smooth envelope after fitting.

[0069] The denoising unit further smooths the envelope. It extracts 16 nearby points from any point in the ordered point set to form a new ordered point set. A quadratic curve is then fitted to these points, and the least squares method is used to solve for the fitting parameters of the quadratic curve, thereby generating a continuous and smooth envelope. In ultrasonic equipment, the denoising unit smooths the calculated envelope, removing jagged edges and abrupt changes, making the envelope more continuous and smooth.

[0070] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows:

[0071] This invention achieves continuous and smooth rendering of the envelope by using quadratic curve fitting based on local neighborhood and denoising technology based on least squares, which significantly improves the smoothness and accuracy of the signal.

[0072] This invention achieves real-time processing and display through efficient data processing algorithms, ensuring the high efficiency and reliability of ultrasound equipment. Attached Figure Description

[0073] Figure 1 This is a schematic flowchart of the method of the present invention;

[0074] Figure 2 This is a schematic diagram of the structure of the ultrasonic device of the present invention;

[0075] Figure 3 This is a schematic diagram of the image data collected in an embodiment of the present invention;

[0076] Figures 4a-4b These are schematic diagrams showing the boundary line determination process before and after processing in an embodiment of the present invention.

[0077] Figures 5a-5b These are schematic diagrams of the main body extracted in this embodiment of the invention before and after processing.

[0078] Figure 6 This is a schematic diagram of the calculated upper and lower edge points in an embodiment of the present invention;

[0079] Figures 7a-7b para in the embodiments of the present invention 3 A diagram illustrating the impact on the results. Figure 7a In Figure 7B, para3 = 100, para 3 =1000;

[0080] Figures 8a-8bThese are schematic diagrams of the envelope before and after quadratic curve fitting in the embodiments of the present invention. Detailed Implementation

[0081] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0082] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0083] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0084] When using expressions such as "at least one of A, B, and C," the expression should generally be interpreted in accordance with the meaning commonly understood by a person skilled in the art (e.g., "a system having at least one of A, B, and C" should include, but is not limited to, systems having A alone, having B alone, having C alone, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.). Similarly, when using expressions such as "at least one of A, B, or C," the expression should generally be interpreted in accordance with the meaning commonly understood by a person skilled in the art (e.g., "a system having at least one of A, B, or C" should include, but is not limited to, systems having A alone, having B alone, having C alone, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.).

[0085] The accompanying drawings illustrate several block diagrams and / or flowcharts. It should be understood that some blocks, or combinations thereof, in the block diagrams and / or flowcharts can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that, when executed by the processor, these instructions can create means for implementing the functions / operations described in these block diagrams and / or flowcharts. The technology of this disclosure can be implemented in hardware and / or software (including firmware, microcode, etc.). Alternatively, the technology of this disclosure can take the form of a computer program product stored on a computer-readable storage medium, which is available for use by or in conjunction with an instruction execution system.

[0086] In PW mode, the ultrasound probe continuously emits ultrasound waves toward a blood vessel in the human body and continuously acquires RF data. Based on the Doppler effect, the frequency information calculated from this RF data can characterize the blood flow velocity in that blood vessel. To facilitate visualization, the frequency information needs to be dynamically compressed (i.e., log compression).

[0087] Assume a total of n image data segments {x1, x2, x3, ... x} were collected. n}, where x i (1≤i≤n) represents the grayscale data formed after log compression of the frequency information of the acquired RF data. It is a two-dimensional real matrix of height×width (height=384, width=16 in the sequence). For each x i (1≤i≤n), the vertical data displays frequency information (or blood flow velocity information), and the horizontal data displays time information. For ease of discussion, assume that the input data has been shifted through the baseline so that the baseline is located at the exact center.

[0088] like Figure 3 As shown, in ultrasound PW mode, continuously acquired x-rays i (1≤i≤n), in chronological order, after splicing the images, the image on the right will be formed. This image is the one that the ultrasound equipment interface needs to display. The PW envelope is a curve drawn in this image, which can characterize the boundary value of blood flow velocity.

[0089] Step 1: Determine the noise distribution

[0090] from Figure 3 It can be seen that x i The image generated by (1≤i≤n) contains a lot of noise. In order to control the impact of noise on the final result, it is necessary to first evaluate the noise distribution function. Generally speaking, the noise in PW mode approximately follows a Gaussian distribution, and x i(1≤i≤n) is the data obtained after log compression. The log compression process involves two parameters, log(D) and log(G), which directly affect the mean and variance of the noise. Assume the noise data ε~N(μ,σ) 2 ), where μ is the mathematical expectation of the noise ε, σ 2 These are the variances. They are affected by the control parameters log(D) and log(G) of the ultrasound system.

[0091] In the system, μ can be approximated using statistical methods, let the current frame be x. n ,

[0092] but

[0093] Where para1 is the time parameter, and para1 = 80 can be selected. In this case, μ is the average value of the accumulated data at past para1 time points, replacing the noise mean. Here, i represents the index in the time direction, and j represents the index in the spatial direction. For example, image data is I(i, j).

[0094] Step 2: Find the dividing line

[0095] The PW envelope is divided into two directions: an upward envelope and a downward envelope. For each x... i For a sequence (1≤i≤n), a dividing line needs to be determined. The portion above the dividing line is used to calculate the upward envelope, and the portion below the dividing line is used to calculate the downward envelope. If the dividing line is not determined, the upward and downward envelopes will intersect and become disordered, affecting the envelope effect.

[0096] In other words, it is necessary to perform a test on each x∈{x1,x2,x3,…x} n Find an integer med∈[1,height] such that med can separate the regions of the upper and lower envelopes in x.

[0097] Find the local average:

[0098]

[0099] Binarization of local averages:

[0100]

[0101] para2 is generally the average difference σ of the noise ε of PW, and can also be taken as 10.

[0102] Generate weight vector:

[0103] Where δ = 160000

[0104] Convolution of the weight vector and the binarized vector:

[0105]

[0106] The maximum value of the convolution is the desired dividing line:

[0107]

[0108] med is the boundary between the upper and lower envelopes.

[0109] Figure 4a The white horizontal lines shown are the dividing lines med we are looking for. Figure 4b This is the processed image.

[0110] The horizontal line found in this way ensures that the distance between the black areas on the top and bottom sides and the horizontal line is as far apart as possible.

[0111] Step 3: Extract the main body

[0112] For each x∈{x1,x2,x3,…x n The signals x and y contain a lot of noise, which interferes with the final envelope output, making the envelope unstable and causing jumps. To remove the noise interference and make the image clean and the signal stable, the main part of the signal needs to be extracted. The expected value of the signal is then calculated by averaging over time and subtracting the noise.

[0113] Where 1≤j≤height, 1≤i≤width

[0114] like Figures 5a-5b As shown, each x∈{x1,x2,x3,…x} n After subtracting noise, the main data is obtained and then stitched together to form an image (i.e., Figure 5b ).

[0115] Step 4: Find the envelope point for each line individually.

[0116] For each x i (1≤i≤n), after finding the boundary line, start calculating the upward envelope. The upward envelope can be subdivided into two lines: the upper edge envelope and the lower edge envelope. (The downward envelope can also be subdivided into two lines, and the calculation method is similar.)

[0117] For each column of data in maindata, calculate an upper edge point and a lower edge point. This process is performed independently for different columns. Specifically, for each i∈{1,2,3,…width}, extract the corresponding column vector SingleLine from maindata.

[0118] SingleLine(j) = MainData(j,i), where 1 ≤ j ≤ height

[0119] Based on the previously calculated boundary lines between the upper and lower envelopes, the upper envelope data is extracted as follows:

[0120] UpEnvelopeData(j) = SingleLine(height-j), where 1≤j≤height-med

[0121] Calculate the sum of the first n terms

[0122] Where 1≤j≤height-med

[0123] Calculate the score: para3 is the adjustment term, which is selected as a constant of 1000 in this example.

[0124] Where 1≤j≤height-med

[0125] Calculate the difference as the score:

[0126] score(j) = Acuu(j) - ExpectIntegral(j), where 1 ≤ j ≤ height - med

[0127]

[0128] like Figure 6 As shown, Figure 6 The green curve represents the score, with the maximum value corresponding to the outer envelope and the minimum value corresponding to the inner envelope.

[0129] like Figures 7a-7b The image shows the effect of para3 on the results: where, Figure 7a It is para3 = 100. Figure 7b It is para3 = 1000. It can be seen that the larger para3 is, the larger the area enclosed by the envelope.

[0130] Step 5: Ordered point set denoising

[0131] from Figures 7a-7b As shown, the envelope displays blood flow velocity information, exhibiting a sawtooth and abrupt changes. In reality, blood flow velocity in human blood vessels changes uniformly and does not exhibit sawtooth or similar phenomena. The sawtooth effect in the calculated velocity values ​​is related to various factors, including noise in the acquired RF data. To eliminate this sawtooth effect, noise reduction processing of the envelope is necessary.

[0132] Let W = width. i(1≤i≤n) will generate an ordered point set {x} i1 ,x i2 ,x i3 ,…,x iW Thus, the image sequence {x1,x2,x3,…x} n This will generate an ordered set of points:

[0133] {x 11 ,x 12 ,x 13 ,…,x 1W ,x 21 ,x 22 ,x 23 ,…,x 2W ,x 31 ,x 32 ,x 33 ,…,x 3W ,x 41 …,x (n-1)W ,x n1 ,x n2 ,x n3 ,…,x nW There are a total of n*width points. Connecting these n*width points in sequence will result in a curve. To eliminate outliers and make the curve smooth, a quadratic curve fitting based on the local neighborhood is needed.

[0134] For ease of discussion, let's denote the ordered point set above as {X1,X2,X3,…,X...} M For an ordered point set {X1, X2, X3, ..., X} M Any point in} (where k is X) k (Indices in this set). Extract 16 nearby points to form an ordered point set {X}. k-7 ,X k-6 ,X k-5 ,…,X k+8}, find a quadratic curve y = ax 2 +bx+c (where a, b, c are the parameters to be calculated) to fit the ordered point set {X} k-7 ,X k-6 ,X k-5 ,…,X k+8 That is, to find variables a, b, c such that the following system of equations holds true.

[0135]

[0136] This is an overdetermined system of equations with variables a, b, and c. Only one least-squares solution can be found. That is:

[0137] Let matrix

[0138] The above overdetermined system of equations can be simplified to A(a,b,c). Τ =y, and its least squares solution is: (a,b,c) Τ =(A Τ A) -1 A Τ β.

[0139] Let x = 8, then the value of the fitted quadratic curve is a*8. 2 +b*8+c.

[0140] like Figures 8a-8b As shown, after quadratic curve fitting, the envelope becomes continuous and smooth, and the local maximum and local minimum points do not become "flat" due to fitting.

[0141] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0142] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0143] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0144] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0145] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0146] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0147] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0148] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0149] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for automatic real-time envelope plotting in PW mode of an ultrasonic device, characterized in that, Includes the following steps: In PW mode, an ultrasound probe is used to emit ultrasound waves to the target vascular tissue and continuously acquire RF data. The collected RF data is log-compressed to form grayscale data, and the expected value and variance of the noise are calculated based on the grayscale data. For each frame of image data, the boundary line is calculated to separate the upper and lower envelopes; Extract the main part of the signal from each frame of image data, subtract the mathematical expectation of the noise, and obtain the processed image data; Calculate the edge points of the upper and lower envelopes for each column of data in each frame of image data; The obtained edge points are subjected to ordered point set denoising processing, and a quadratic curve fitting method based on local neighborhood is used to generate a continuous and smooth envelope.

2. The automatic real-time envelope drawing method for ultrasound equipment in PW mode according to claim 1, characterized in that, The grayscale data is subjected to baseline shifting so that the baseline is located in the center of the image.

3. The automatic real-time envelope drawing method for ultrasound equipment in PW mode according to claim 1, characterized in that, The calculation steps for the boundary line include: Calculate the local average value; Binarize the local average value; Generate a weight vector; Perform convolution on the binary vector and the weight vector, and use the maximum value of the convolution as the boundary.

4. The automatic real-time envelope drawing method for ultrasound equipment in PW mode according to claim 1, characterized in that, The main body extraction step of the signal includes: Averaging of image data along the time axis; The mathematical expectation of the noise is subtracted to obtain the processed image data.

5. The automatic real-time envelope drawing method for ultrasound equipment in PW mode according to claim 1, characterized in that, The steps for calculating the edge points of the envelope include: Extract the upper and lower envelope portions of the data based on the dividing line; Calculate the sum of the first n terms of each column of data; Calculate the score, select the maximum value as the outer envelope, and the minimum value as the inner envelope.

6. The automatic real-time envelope drawing method for ultrasound equipment in PW mode according to claim 1, characterized in that, The ordered point set denoising process includes: For any point in an ordered set of points, extract 16 points from its vicinity to form another ordered set of points. Find a quadratic curve to fit an ordered set of points; The least squares method is used to solve for the fitting parameters of the quadratic curve; Generate a continuous and smooth envelope after fitting.

7. An ultrasonic device, characterized in that, include: An ultrasound probe is used to emit ultrasound waves to target vascular tissue and continuously acquire RF data in PW mode; The data processing module, connected to the ultrasound probe, is used to perform log compression and baseline shifting on the acquired RF data so that the baseline is located in the center of the image, forming grayscale data. A noise assessment module, connected to the data processing module, is used to calculate the mathematical expectation and variance of the noise based on the grayscale data. The boundary line calculation module is connected to the noise evaluation module and is used to calculate the boundary line for each frame of image data to separate the upper and lower envelopes. The signal extraction module, connected to the boundary line calculation module, is used to extract the main part of the signal from each frame of image data, subtract the mathematical expectation of the noise, and obtain the processed image data. The envelope calculation module, connected to the signal extraction module, is used to calculate the edge points of the upper and lower envelopes for each column of data in each frame of image data, and to perform ordered point set denoising processing on the obtained edge points. A continuous and smooth envelope is generated by using a quadratic curve fitting method based on local neighborhood.

8. The ultrasonic device according to claim 7, characterized in that, The boundary line calculation module includes: Local average value calculation unit, used to calculate local average values; The binarization unit is connected to the local average value calculation unit and is used to binarize the local average value. A weight vector generation unit, connected to the binarization unit, is used to generate weight vectors; The convolution calculation unit, connected to the weight vector generation unit, is used to perform convolution on the binary vector and the weight vector, with the maximum value of the convolution serving as the boundary line.

9. The ultrasonic device according to claim 7, characterized in that, The signal extraction module includes: The average value calculation unit is used to average the image data over time. The noise subtraction unit, connected to the average value calculation unit, is used to subtract the mathematical expectation of the noise to obtain the processed image data.

10. The ultrasonic device according to claim 7, characterized in that, The envelope calculation module includes: The data extraction unit is used to extract the upper envelope and lower envelope portions of the data based on the boundary line; The cumulative sum calculation unit, connected to the data extraction unit, is used to calculate the sum of the first n terms of each column of data; The score calculation unit, connected to the cumulative sum calculation unit, is used to calculate the score, selecting the maximum value point as the outer envelope and the minimum value point as the inner envelope; The noise reduction unit is used to extract 16 nearby points from any point in the ordered point set to form an ordered point set, find a quadratic curve to fit the ordered point set, and use the least squares method to solve the fitting parameters of the quadratic curve to generate a continuous and smooth envelope after fitting.

Images