A three-dimensional ultrasonic shear wave elastic imaging method based on linear array

Abstract

The present invention discloses a three-dimensional ultrasonic shear wave elastic imaging method based on a linear array, comprising the following steps: controlling a linear array ultrasonic sensor to perform free scanning, collecting and saving ultrasonic echo signals, and obtaining a series of two-dimensional ultrasonic elastic images by calculating the shear modulus; combining optical positioning, placing the two-dimensional image at the correct position in the three-dimensional volume, and reconstructing the target tissue in three-dimensional space based on volume data; and utilizing three-dimensional visualization and rendering technology to display the reconstructed three-dimensional graphics reflecting the quantitative elastic information of the target. The present invention takes into account the propagation characteristics of shear waves in three-dimensional space, and the expansion from two-dimensional to three-dimensional ultrasonic shear wave elastic imaging can more accurately measure the shear wave velocity and quantify tissue stiffness, while also providing a more comprehensive physical property and target evaluation from a three-dimensional perspective.

Description

Technical Field

[0001] The invention relates to a shear wave elastic imaging method, in particular to a three-dimensional ultrasonic shear wave elastic imaging method based on a linear array. Background Art

[0002] Shear wave elastography is an ultrasound technology that uses the correlation between tissue elastic modulus and biological properties to evaluate target hardness. It generates transversely propagating shear waves by emitting focused pulse signals, and uses fast plane wave excitation to track the shear wave displacement and velocity, thereby calculating the shear modulus of the tissue.

[0003] By determining mechanical parameters and elastic tissue properties, ultrasonic shear wave elastography supplements the diagnosis of B-mode ultrasound and color Doppler ultrasound, and can examine abnormal information of musculoskeletal soft tissues in early stage lesions, making up for the shortcomings of traditional medical imaging modalities. However, two-dimensional ultrasonic elastic images cannot directly reflect the three-dimensional geometric shape of the target. The diagnostic information it reflects is limited to the two-dimensional cross-sectional information of the tissue, lacks the observation of macroscopic physical properties from a three-dimensional perspective, and the consistency and accuracy between observers need to be improved; considering the characteristics of shear wave propagation in three dimensions, the expansion from two-dimensional to three-dimensional shear wave elastography can promote more accurate shear wave velocity estimation and better quantification of tissue stiffness, thereby providing a more comprehensive target assessment. Summary of the invention

[0004] Purpose of the invention: The technical problem to be solved by the present invention is to provide a three-dimensional ultrasonic shear wave elastic imaging method based on a linear array in view of the shortcomings of the prior art.

[0005] In order to solve the above technical problems, the present invention discloses a three-dimensional ultrasonic shear wave elastic imaging method based on a linear array, comprising the following steps:

[0006] Step 1, transmitting and receiving plane wave echo signals through a linear array ultrasonic sensor, acquiring a B-mode ultrasonic image of the target as a reference image, and determining a region of interest for ultrasonic shear wave elastography;

[0007] Step 2, emitting ultrasonic signals with different time delays through the sensor units of the linear array ultrasonic sensor, focusing the ultrasonic beam at the target depth and position, and generating transversely propagating shear waves; using fast plane wave excitation to track the displacement and velocity of the shear wave during propagation, calculating the tissue displacement through the speckle tracking algorithm, calculating the shear wave velocity and shear modulus, and combining color coding technology to obtain a two-dimensional ultrasonic quantitative elastic image;

[0008] Step 3, controlling the linear array ultrasonic sensor to perform free scanning in sequence, generating a two-dimensional ultrasonic quantitative elastic image according to the method described in step 2 for each acquired echo signal, combining with an optical positioning camera to obtain spatial position information, placing each pixel in the two-dimensional image in a three-dimensional volume at the correct position, and reconstructing a target image in three-dimensional space;

[0009] Step 4: Through 3D visualization and rendering technology, volume rendering is used to extract information from the 3D image, and the entire 3D image is projected onto a 2D plane to complete the evaluation of the target physical properties from a 3D perspective.

[0010] In step 1 of the present invention, the linear array ultrasonic sensor includes a group of linearly arranged ultrasonic sensor elements, the positions of all ultrasonic sensor elements are relatively fixed, the distance between each ultrasonic sensor element is the same, and they work independently;

[0011] The method for obtaining the B-type ultrasound image of the target in step 1 is a delayed superposition algorithm. The ultrasound sensor element simultaneously transmits planar ultrasound and receives echo signals. The echo signals received by each ultrasound sensor element are superimposed and summed to obtain the structural information of the target area, determine the area of ​​interest of shear wave elastic imaging, and provide a reference image for the shear wave tracking stage.

[0012] In step 2 of the present invention, the method for focusing the ultrasonic signal is based on the beamforming method of the linear array ultrasonic sensor, comprising:

[0013] By setting each ultrasonic sensor element to transmit ultrasonic signals with different time delays, the wavefront of the sound waves emitted by each ultrasonic sensor element propagates in space and gradually converges to a point, and the beam is focused at the target position; the beam focusing point absorbs the sound waves to generate acoustic radiation force, which is focused at the target position to induce acoustic radiation force, causing the tissue particles at the location to vibrate and generate shear waves that propagate laterally;

[0014] After the ultrasonic beam is focused, a fast plane wave is emitted and the echo signal is received to generate multiple frames of B-mode ultrasound images. The speckle tracking algorithm is used to select matching blocks and search boxes on the ultrasound image, determine the block matching criteria and search method, search and determine the best matching position of the speckle image near the corresponding position of the next frame of the image, that is, the best matching block, and calculate the tissue displacement and shear wave velocity.

[0015] The block matching criterion of the speckle tracking algorithm described in step 2 is the minimum absolute error sum, the absolute error sum ρ m,n The calculation method is as follows:

[0016]

[0017] Among them, X and Y represent two frames of images. If the image is regarded as a two-dimensional matrix of its pixel value position, then Xi,j Represents the pixel value at position (i, j) in image X, Y i+m,j+n represents the pixel value of the search area in image Y, (m,n) represents the displacement of the selected target image block between two frames, l and k are the sizes of the target block; when the absolute error and ρ m,n When it is the smallest, it means that the two image blocks are the best match.

[0018] The search method in the speckle tracking algorithm in step 2 is a diamond search algorithm, which uses two search templates: a small template SDSP with 5 matching points and a large template LDSP with 9 matching points. The method includes:

[0019] Select a region block to be matched in the reference image, set the position of W0 in the region block as the center point, select the large template LDSP to calculate the minimum absolute error sum of the 9 points in the search area including the center point for the region block to be matched, and determine the position of the minimum error block;

[0020] If the minimum error block is not in the center of the search area, the minimum error block is used as the new center point and the above process is repeated until the minimum error block is in the center of the search area.

[0021] The small template SDSP is used to search according to the above steps, and the minimum error block finally obtained is the best matching block.

[0022] The method for calculating the shear modulus described in step 2 includes:

[0023] The shear displacement is calculated between the best matching block and the area block to be matched in the reference image, and the shear modulus μ(x, z) is obtained by calculation. The calculation method includes:

[0024]

[0025] Where μ is the shear modulus at the pixel coordinate (x, z). The flat scan method approximates that there is no obvious diffraction in the y direction, and the shear displacement at the pixel coordinate (x, z) on the xz plane is u z (x,z), t is time, ρ is medium density, ω is center frequency, N is the number of frequencies, is a threshold, and F is Fourier transform in time domain. is a second-order partial derivative operation; after the shear modulus is calculated according to the formula, the Young's modulus E is calculated by the formula E≈3μ, and a two-dimensional ultrasonic quantitative elastic image is obtained.

[0026] In step 2 of the present invention, the color coding technology converts the grayscale of the two-dimensional ultrasonic elastic image into a color image through a coding method, wherein the image part with a small elastic coefficient is encoded as red, indicating that the corresponding tissue is soft; the image part with a large elastic coefficient is encoded as blue, indicating that the corresponding tissue is hard; the image part with a medium elastic coefficient is encoded as green.

[0027] In step 3 of the present invention, the target graphics of the three-dimensional space are reconstructed based on volume data, comprising:

[0028] Step 3-1, construct the data structure:

[0029] Determine the size of the reconstructed volume data according to the information of the two-dimensional ultrasound elastic image, the size scale includes the size of the volume data dimension, the coordinate origin information and the space interval between voxels, and determine the size of the reconstructed volume data by a bounding box-based technology;

[0030] Step 3-2, reallocate sampling pixels:

[0031] Traverse each pixel point on the two-dimensional elastic image, rearrange the position information of the pixel points through the transformation matrix, and map the pixel to the correct position of the three-dimensional volume data. The mapping method is forward mapping, including:

[0032] u i =MX i (3)

[0033] Among them, u i represents the corresponding voxel coordinates in the reconstructed volume data, M is the forward transformation matrix, X i is the pixel coordinate on the two-dimensional ultrasound quantitative elasticity image;

[0034] Step 3-3, calculate the empty voxel value in the volume data:

[0035] After the pixels are reallocated as described in step 3-2, there are blank areas in the reconstructed volume, and the unknown voxel values ​​in the voxel grid are supplemented by interpolating the surrounding known pixel values.

[0036] The algorithm used in the volume data reconstruction method in step 3 of the present invention is a pixel-based method, which is implemented by reallocating pixels and filling blank area voxels. The pixel reallocation stage adopts the nearest pixel assignment method, which assigns the nearest pixel value found after traversal to the voxel according to the forward mapping. If the voxel value is originally empty, it is directly assigned. If it is not empty, multiple pixel values ​​corresponding to the voxel are averaged and then assigned.

[0037] The stage of filling empty voxels described in step 3-3 adopts the traditional distance weighted algorithm, which is as follows:

[0038]

[0039] Among them, traverse the empty voxel points u that need to be filled j , with the voxel point as the center and r as the radius, determine a circular neighborhood N i , j∈N j is a set of known voxels in the neighborhood, according to the distance weighted rule ω j The weighted average of these known voxels is used to assign values ​​to the corresponding empty voxel points.

[0040] In step 4 of the present invention, the method used by the three-dimensional visualization rendering technology is volume-based rendering, which uses rays to project a three-dimensional image and project it onto a plane for display; the voxels along each ray are weighted according to the following method:

[0041] C out =C in [1-a(j)]+c j a j (5)

[0042] Among them, C out is the value of the ray coming out of the jth voxel, C in is the value of the ray entering the jth voxel, parameter a controls the opacity, parameter c j It is a parameter adjusted based on the local gradient, controlling the brightness of the voxel;

[0043] When a(j) = 0, the voxel is transparent and the ray is emitted through the jth voxel; when a(j) = 1, the parameter c j Determines whether the voxel is opaque or emissive.

[0044] Beneficial effects:

[0045] Considering the three-dimensional propagation characteristics of ultrasonic shear waves in tissues, reconstructing a spatial stereogram reflecting the elastic information of diseased tissues can promote more accurate shear wave velocity estimation and tissue stiffness quantification.

[0046] Integrating three-dimensional ultrasound technology into ultrasonic shear wave elastography technology overcomes the viewing angle limitation of two-dimensional ultrasound elastography, expands the image information originally limited to two-dimensional cross-sections to a global three-dimensional perspective, provides more intuitive multi-level tissue spatial information, ensures consistency and accuracy of observations among doctors, and effectively improves the efficiency of doctors in locating lesions and assessing their grades.

[0047] The free scanning sensor scanning method gives the operator a relatively free collection method, improving the characteristics of two-dimensional ultrasonic elastic imaging that is sensitive to the pressure and scanning angle of the sensor. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more clear.

[0049] Figure 1 is a flow chart of three-dimensional elastic image reconstruction and display of the present invention;

[0050] Figure 2 It is a schematic diagram of reconstruction of a two-dimensional shear wave elastic image based on a linear array according to the present invention. DETAILED DESCRIPTION

[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0052] A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array comprises the following steps:

[0053] Step 1, using a linear array ultrasonic sensor to transmit and receive a plane wave echo signal, obtaining a B-mode ultrasonic image of the target as a reference image, and determining a region of interest (ROI) for ultrasonic shear wave elastography;

[0054] The linear array ultrasonic sensor comprises a group of ultrasonic sensor elements arranged linearly, the positions of all ultrasonic sensor elements are relatively fixed, the distance between each ultrasonic sensor element is the same, and they work independently;

[0055] The method for acquiring the B-type ultrasound image of the target is a delay-and-sum beamforming algorithm (reference: Klemm, M., Craddock, IJ, Leendertz, JA, Preece, A., & Benjamin, R. (2008). Improved delay-and-sum beamforming algorithm for breast cancer detection. International Journal of Antennas and Propagation, 2008.), wherein the ultrasound sensor element simultaneously transmits planar ultrasound and receives echo signals, and the echo signals received by each ultrasound sensor element are superimposed and summed to acquire the structural information of the target area, determine the region of interest of shear wave elastography, and provide a reference image for the shear wave tracking stage.

[0056] Step 2, emitting ultrasonic signals with different time delays through the sensor units of the linear array ultrasonic sensor, focusing the ultrasonic beam at the target depth and position, and generating transversely propagating shear waves; using fast plane wave excitation to track the displacement and velocity of the shear wave during propagation, calculating the tissue displacement through the speckle tracking algorithm, calculating the shear wave velocity and shear modulus, and combining color coding technology to obtain a two-dimensional ultrasonic quantitative elastic image;

[0057] The method for focusing the ultrasonic signal is based on the beamforming method of the linear array ultrasonic sensor, comprising:

[0058] By setting each ultrasonic sensor element to transmit ultrasonic signals with different time delays, the wavefront of the sound waves emitted by each ultrasonic sensor element propagates in space and gradually converges to a point, and the beam is focused at the target position; the beam focusing point absorbs the sound waves to generate acoustic radiation force, which is focused at the target position to induce acoustic radiation force, causing the tissue particles at the location to vibrate and generate shear waves that propagate laterally;

[0059] After the ultrasonic beam is focused, a fast plane wave is emitted and the echo signal is received to generate multiple frames of B-mode ultrasound images. The speckle tracking algorithm is used to select matching blocks and search boxes on the ultrasound image, determine the block matching criteria and search method, search and determine the best matching position of the speckle image near the corresponding position of the next frame of the image, that is, the best matching block, and calculate the tissue displacement and shear wave velocity.

[0060] The block matching criterion of the speckle tracking algorithm is the minimum absolute difference (SAD, Sum of Absolute Differences), the absolute difference and ρ m,n The calculation method is as follows:

[0061]

[0062] Among them, X and Y represent two frames of images. If the image is regarded as a two-dimensional matrix of its pixel value position, then X i,j Represents the pixel value at position (i, j) in image X, Y i+m,j+n represents the pixel value of the search area in image Y, (m,n) represents the displacement of the selected target image block between two frames, l and k are the sizes of the target block; when the absolute error and ρ m,n When it is the smallest, it means that the two image blocks are the best match.

[0063] The search method in the speckle tracking algorithm is a diamond search algorithm, which adopts two search templates: a small template SDSP (SDSP, Small Diamond Search Pattern) with 5 matching points and a large template LDSP (LDSP, Large Diamond Search Pattern) with 9 matching points. The method includes:

[0064] Select a region block to be matched in the reference image, set the position of W0 in the region block as the center point, select the large template LDSP to calculate the minimum absolute error sum of the 9 points in the search area including the center point for the region block to be matched, and determine the position of the minimum error block;

[0065] If the minimum error block is not in the center of the search area, the minimum error block is used as the new center point and the above process is repeated until the minimum error block is in the center of the search area.

[0066] The small template SDSP is used to search according to the above steps, and the minimum error block finally obtained is the best matching block.

[0067] The method for calculating the shear modulus comprises:

[0068] The shear displacement is calculated between the best matching block and the area block to be matched in the reference image, and the shear modulus μ(x, z) is obtained by calculation. The calculation method includes:

[0069]

[0070] Where μ is the shear modulus at the pixel coordinate (x, z). The flat scan method approximates that there is no obvious diffraction in the y direction, and the shear displacement at the pixel coordinate (x, z) on the xz plane is u z (x,z), t is time, ρ is medium density, ω is center frequency, N is the number of frequencies, is a threshold, and F is Fourier transform in time domain. is a second-order partial derivative operation; after the shear modulus is calculated according to the formula, the Young's modulus E is calculated by the formula E≈3μ, and a two-dimensional ultrasonic quantitative elastic image is obtained.

[0071] The color coding technology converts the grayscale of a two-dimensional ultrasonic elastic image into a color image through a coding method, wherein the image portion with a small elastic coefficient is encoded as red, indicating that the corresponding tissue is soft; the image portion with a large elastic coefficient is encoded as blue, indicating that the corresponding tissue is hard; and the image portion with a medium elastic coefficient is encoded as green.

[0072] Step 3, controlling the linear array ultrasonic sensor to perform free scanning in sequence, generating a two-dimensional ultrasonic quantitative elastic image according to the method described in step 2 for each acquired echo signal, combining with an optical positioning camera to obtain spatial position information, placing each pixel in the two-dimensional image in a three-dimensional volume at the correct position, and reconstructing a target image in three-dimensional space;

[0073] The target graphics of the reconstructed three-dimensional space is reconstructed based on volume data, comprising:

[0074] Step 3-1, construct the data structure:

[0075] The size of the reconstructed volume data is determined according to the information of the two-dimensional ultrasound elastic image. The size includes the size of the volume data dimension, the coordinate origin information and the spatial interval between voxels. The size of the reconstructed volume data is determined by a bounding box-based technology (reference: Fan Yangyang, Ni Qian, Wen Tiexiang, Gu Jia, Wang Lei, Xie Yaoqin, & Liu Jingen. (2017). The latest research progress of Freehand three-dimensional ultrasound volume data reconstruction. Chinese Journal of Biomedical Engineering, 36(1), 92-102.);

[0076] Step 3-2, reallocate sampling pixels:

[0077] Traverse each pixel point on the two-dimensional elastic image, rearrange the position information of the pixel points through the transformation matrix, and map the pixel to the correct position of the three-dimensional volume data. The mapping method is forward mapping, including:

[0078] u i =MX i (3)

[0079] Among them, u i represents the corresponding voxel coordinates in the reconstructed volume data, M is the forward transformation matrix, X i is the pixel coordinate on the two-dimensional ultrasound quantitative elasticity image;

[0080] Step 3-3, calculate the empty voxel value in the volume data:

[0081] After the pixels are reallocated as described in step 3-2, there are blank areas in the reconstructed volume, and the unknown voxel values ​​in the voxel grid are supplemented by interpolating the surrounding known pixel values.

[0082] The algorithm adopted by the volume data reconstruction method is a pixel-based method (PBM), which is realized by reallocating pixels and filling blank area voxels. The pixel reallocation stage adopts the nearest pixel assignment method, which assigns the nearest pixel value found after traversal to the voxel according to the forward mapping. If the voxel value is originally empty, it is directly assigned. If it is not empty, multiple pixel values ​​corresponding to the voxel are averaged and then assigned.

[0083] The stage of filling empty voxels described in step 3-3 adopts the traditional distance weighted algorithm, which is as follows:

[0084]

[0085] Among them, traverse the empty voxel points u that need to be filled j , with the voxel point as the center and r as the radius, determine a circular neighborhood Ni , j∈N j is a set of known voxels in the neighborhood, according to the distance weighted rule ω j The weighted average of these known voxels is used to assign values ​​to the corresponding empty voxel points.

[0086] Step 4: Through 3D visualization and rendering technology, volume rendering is used to extract information from the 3D image, and the entire 3D image is projected onto a 2D plane to complete the evaluation of the target physical properties from a 3D perspective.

[0087] The method used in the three-dimensional visualization rendering technology is volume-based rendering, which uses rays to project a three-dimensional image and project it onto a plane for display; the voxels along each ray are weighted according to the following method:

[0088] C out =C in [1-a(j)]+c j a j (5)

[0089] Among them, C out is the value of the ray coming out of the jth voxel, C in is the value of the ray entering the jth voxel, parameter a controls the opacity, parameter c j It is a parameter adjusted based on the local gradient, controlling the brightness of the voxel;

[0090] When a(j) = 0, the voxel is transparent and the ray is emitted through the jth voxel; when a(j) = 1, the parameter c j Determines whether the voxel is opaque or emissive.

[0091] Example

[0092] An embodiment of the present invention discloses a three-dimensional ultrasonic shear wave elastic imaging method based on a linear array. The method adopts a free-scanning one-dimensional linear array probe to obtain a two-dimensional ultrasonic elastic image, reconstructs the image sequence to obtain three-dimensional image volume data for rendering and display, thereby improving the accuracy of shear wave velocity measurement and tissue stiffness quantification, and at the same time provides richer three-dimensional spatial structural information, further improving the consistency of observation of the target physical properties and the accuracy of evaluation.

[0093] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0094] like Figure 1 As shown, an embodiment of the present invention provides a three-dimensional ultrasonic shear wave elastic imaging method based on a linear array, comprising:

[0095] Step 1, transmitting and receiving plane wave echo signals through a linear array ultrasonic sensor, acquiring a B-mode ultrasonic image of the target as a reference image, and determining a region of interest for ultrasonic shear wave elastography;

[0096] Step 2, emitting ultrasonic signals with different time delays through the sensor units of the linear array ultrasonic sensor, so that the ultrasonic beam is focused on the target depth and position, generating a shear wave that propagates laterally; using fast plane wave excitation to track the displacement and velocity of the shear wave during propagation, calculating the tissue displacement through a speckle tracking algorithm, and then calculating the shear wave velocity and shear modulus, combined with color coding technology, thereby obtaining a two-dimensional ultrasonic quantitative elastic image;

[0097] Step 3, controlling the linear array ultrasonic sensor to perform free scanning in a certain order, generating a two-dimensional ultrasonic elastic image according to the above steps for each acquired echo signal, combining with an optical positioning camera to obtain spatial position information, placing each pixel in the two-dimensional image in a three-dimensional volume at the correct position, and reconstructing a target image in the three-dimensional space;

[0098] Step 4: Through 3D visualization and rendering technology, volume rendering is used to extract important information from the 3D image, and the entire 3D image is projected onto a 2D plane for display, so as to achieve the target physical property evaluation from a 3D perspective.

[0099] The ultrasonic sensor elements of the linear array described in this embodiment are arranged linearly, and their specific position distribution is related to the sensor model used. The distance between each ultrasonic sensor element is the same, and they can work independently; in the plane wave transmission and acquisition stage, all ultrasonic sensor elements are turned on and working, and in the delayed focusing stage, only some ultrasonic sensor elements are turned on to transmit ultrasonic signals, so that the focused beam is more concentrated to ensure a better focusing effect.

[0100] like Figure 2 As shown, the experimental equipment of this embodiment is a Verasonics Vantag64 system equipped with an L7-4 ultrasonic sensor probe, with 128 ultrasonic sensor elements, the interval between each ultrasonic sensor element is about 3mm, the center frequency of the transmitted ultrasonic signal is 5.208Mhz, and the background sound speed in the tissue is set to 1540m / s. In the implementation process, 64 ultrasonic sensor elements are used in the plane wave transmission and acquisition stage, 32 ultrasonic sensor elements are used in the focusing process, and the focusing depth is set to 5cm.

[0101] The reconstruction method of the B-type ultrasound image described in this embodiment is a time-delay superposition algorithm. The image is acquired by transmitting and receiving a plane wave, and is used as a reference image in the focusing stage. The focusing depth and the scanning area of ​​the sensor are set according to the region of interest, and then time-delay focusing is performed. The specific implementation method of the focusing delay is the center focusing of the sound beam. Assuming that the number of ultrasonic sensor elements in the linear array is H, and the distance between each sensor element is d, then the distance between the i-th sensor element in the sequence and the set center sensor element (the j-th sensor element, directly above the focus) is |ij|d; the distance from the center sensor element to the focus on the center line is set to D, then the distance D between the i-th ultrasonic sensor element and the focus is i It can be obtained by calculating the Pythagorean theorem. Assume that the kth sensor element is the farthest from the central sensor element. Take the emission time of the farthest kth sensor element as the benchmark. If the speed of sound is known to be c, then the delay value required by the i-th sensor element is Δt i , the calculation formula is as follows:

[0102]

[0103]

[0104]

[0105] By setting different delayed excitations for each ultrasonic sensor element according to the above process, central focusing can be achieved at the target depth.

[0106] In this embodiment, after the ultrasonic beam is focused, the linear array ultrasonic sensor immediately transmits a fast plane wave and receives the echo signal to generate a multi-frame B-mode ultrasonic image, and the speckle tracking algorithm is used to calculate the tissue displacement and shear wave velocity. The matching block size selected in the speckle tracking algorithm is set to 3×3 (pixels), and the search window is set to 10×10 (pixels); the block matching criteria are the minimum absolute error and SAD, and the definition formula is as follows:

[0107]

[0108] Where X i,j It represents the pixel value of (i, j) in image X, Y i+m,j+n represents the pixel value of the search area in image Y, (m,n) represents the displacement of the selected target image block between the two frames, l and k are the sizes of the target block, and when the absolute error and ρ m,nWhen the error is the smallest, the two image blocks are considered to be the most matched; the search method used by speckle tracking is the diamond search algorithm, which adopts two search templates, namely the small template SDSP with 5 matching points and the large template LDSP with 9 matching points. The algorithm flow is as follows: first, the position of W0 in the matching block of the reference image is set as the center point, and the LDSP is used to calculate the SAD values ​​of the 9 points including the center point in the search area of ​​the matching image (between the reference image and the matching image), and the position of the minimum error block is determined. Secondly, it is determined whether the minimum error block is in the center of the search area. If so, the minimum error block is used as the new center point, and the above process is repeated until the minimum error block is located in the center of the search area. Finally, the SDSP is used to search according to the above steps, and the minimum error block obtained is the best matching block.

[0109] In this embodiment, after the shear displacement between the best matching block and the reference image matching block is obtained, the shear modulus can be further calculated by the formula:

[0110]

[0111] Where μ is the shear modulus. Since the flat scanning method can be considered without obvious diffraction in the y direction, only the shear displacement u on the xz plane needs to be considered. z (x,z), ρ is the medium density, ω is the center frequency, N represents the frequency number, that is, a threshold is selected in the Laplace spectrum near the center frequency of the shear wave, and F represents the Fourier transform in the time domain. Young's modulus E can also be calculated by the formula E≈3μ, thereby obtaining a two-dimensional grayscale quantitative elastic image. The two-dimensional elastic image in the previous step is converted from grayscale to color image through color coding technology to improve the effect of visual interpretation of the image.

[0112] The reconstruction method of the three-dimensional image adopted in this embodiment is a reconstruction method based on volume data. The reconstruction steps are divided into three steps: constructing a volume data structure, reallocating sampled pixels, and calculating voxel values ​​in the volume data. The algorithm adopted by the reconstruction method based on volume data is a pixel-based method PBM, which is specifically implemented by reallocating pixels and filling voxels in blank areas. In order to improve the interpolation effect, the empty voxel filling stage can be optimized from the traditional distance weighted algorithm to the fast marching method (FMM), where the interpolation formula of the empty voxels is as follows:

[0113]

[0114]

[0115] Consider a region to be filled in the reconstructed voxel value, p is the voxel on the boundary (step boundary) of the region to be filled, B ε(p) is the spherical interpolation region with p as the center and ε as the radius, q∈B ε (p) represents the known voxels in the region, d pq is the distance between voxels p and q, w(p,q) is the weight between voxels p and q, n is the normal vector of the step boundary, v is the vector from voxel p to q, is the gradient value at point q, and u(p) represents the interpolation of known voxels in a spherical area with p as the center and ε as the radius.

[0116] The 3D visualization rendering technology method used in this embodiment is based on volume rendering, which is based on the principle of using rays to project a 3D image onto a plane for display. In this volume rendering technology, the voxels along each ray are weighted according to the following formula:

[0117] C out =C in [1-a(j)]+c j a j (5)

[0118] Among them C out is the value of the ray coming out of the jth voxel, C in is the value of the ray entering the jth voxel. Parameter a is used to control opacity, and parameter c j It is a parameter adjusted based on the local gradient, or used to control the brightness of the voxel. When a(j) = 0, the voxel is considered transparent and the ray is emitted through the jth voxel; when a(j) = 1, the parameter c j To determine whether a voxel is opaque or emissive.

[0119] Compared with the prior art, the present invention has the following advantages: taking into account the three-dimensional propagation characteristics of ultrasonic shear waves in tissues, reconstructing a spatial stereogram reflecting the elastic information of the lesion target can promote more accurate shear wave velocity estimation and tissue stiffness quantification; integrating three-dimensional ultrasound technology into shear wave elastic imaging technology overcomes the viewing angle limitation of two-dimensional ultrasonic elastic imaging, and expands the image information originally limited to two-dimensional cross-sections to a global three-dimensional perspective, providing more intuitive multi-level tissue spatial information, ensuring the consistency and accuracy of observations among doctors, and effectively improving the efficiency of doctors in locating lesion positions and evaluating grades; the free scanning sensor scanning method gives the operator a relatively free collection method, improving the characteristics of two-dimensional ultrasonic elastic imaging that is sensitive to sensor pressure and scanning angle.

[0120] The present invention provides a concept and method of a three-dimensional ultrasonic shear wave elastic imaging method based on a linear array. There are many methods and ways to implement the technical solution. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention. All components not specified in this embodiment can be implemented using existing technologies.

Claims

1. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array, characterized in that: The following steps are involved: Step 1, transmitting and receiving plane wave echo signals through a linear array ultrasonic sensor, acquiring a B-mode ultrasonic image of the target as a reference image, and determining a region of interest ROI for ultrasonic shear wave elastography; Step 2, emitting ultrasonic signals with different time delays through the sensor units of the linear array ultrasonic sensor, focusing the ultrasonic beam at the target depth and position, and generating transversely propagating shear waves; using fast plane wave excitation to track the displacement and velocity of the shear wave during propagation, calculating the tissue displacement through the speckle tracking algorithm, calculating the shear wave velocity and shear modulus, and combining color coding technology to obtain a two-dimensional ultrasonic quantitative elastic image; Step 3, controlling the linear array ultrasonic sensor to perform free scanning in sequence, generating a two-dimensional ultrasonic quantitative elastic image according to the method described in step 2 for each acquired echo signal, combining with an optical positioning camera to obtain spatial position information, placing each pixel in the two-dimensional image in a three-dimensional volume at the correct position, and reconstructing a target image in three-dimensional space; Step 4: Through 3D visualization and rendering technology, volume rendering is used to extract information from the 3D image, and the entire 3D image is projected onto a 2D plane to complete the evaluation of the target physical properties from a 3D perspective.

2. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 1, characterized in that: In step 1, the linear array ultrasonic sensor includes a group of linearly arranged ultrasonic sensor elements, the positions of all ultrasonic sensor elements are relatively fixed, the distance between each ultrasonic sensor element is the same, and they work independently; The method for obtaining the B-type ultrasound image of the target in step 1 is a delayed superposition algorithm. The ultrasound sensor element simultaneously transmits planar ultrasound and receives echo signals. The echo signals received by each ultrasound sensor element are superimposed and summed to obtain the structural information of the target area, thereby determining the area of ​​interest of shear wave elastic imaging and providing a reference image for the shear wave tracking stage.

3. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 2, characterized in that: In step 2, the method for focusing the ultrasonic signal is based on the beamforming method of the linear array ultrasonic sensor, including: By setting each ultrasonic sensor element to transmit ultrasonic signals with different time delays, the wavefront of the sound waves emitted by each ultrasonic sensor element propagates in space and gradually converges to a point, and the beam is focused at the target position; the beam focusing point absorbs the sound waves to generate acoustic radiation force, which is focused at the target position to induce acoustic radiation force, causing the tissue particles at the location to vibrate and generate shear waves that propagate laterally; After the ultrasonic beam is focused, a fast plane wave is emitted and the echo signal is received to generate multiple frames of B-mode ultrasound images. The speckle tracking algorithm is used to select matching blocks and search boxes on the ultrasound image, determine the block matching criteria and search method, search and determine the best matching position of the speckle image near the corresponding position of the next frame of the image, that is, the best matching block, and calculate the tissue displacement and shear wave velocity.

4. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 3, characterized in that: The block matching criterion of the speckle tracking algorithm described in step 2 is the minimum absolute error sum, the absolute error sum ρ m,n The calculation method is as follows: Among them, X and Y represent two frames of images. If the image is regarded as a two-dimensional matrix of its pixel value position, then X i,j Represents the pixel value at position (i, j) in image X, Y i+m,j+n represents the pixel value of the search area in image Y, (m,n) represents the displacement of the selected target image block between two frames, l and k are the sizes of the target block; when the absolute error and ρ m,n When it is the smallest, it means that the two image blocks are the best match.

5. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 4, characterized in that: The search method in the speckle tracking algorithm in step 2 is a diamond search algorithm, which uses two search templates: a small template SDSP with 5 matching points and a large template LDSP with 9 matching points. The method includes: Select a region block to be matched in the reference image, set the position of W0 in the region block as the center point, select the large template LDSP to calculate the minimum absolute error sum of the 9 points in the search area including the center point for the region block to be matched, and determine the position of the minimum error block; If the minimum error block is not in the center of the search area, the minimum error block is used as the new center point and the above process is repeated until the minimum error block is in the center of the search area. The small template SDSP is used to search according to the above steps, and the minimum error block finally obtained is the best matching block.

6. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 5, characterized in that: The method for calculating the shear modulus described in step 2 includes: The shear displacement is calculated between the best matching block and the area block to be matched in the reference image, and the shear modulus μ(x, z) is obtained by calculation. The calculation method includes: Where μ is the shear modulus at the pixel coordinate (x, z). The flat scan method approximates that there is no obvious diffraction in the y direction, and the shear displacement at the pixel coordinate (x, z) on the xz plane is u z (x,z), t is time, ρ is medium density, ω is center frequency, N is the number of frequencies, is a threshold, and F is Fourier transform in time domain. is a second-order partial derivative operation; after the shear modulus is calculated according to the formula, the Young's modulus E is calculated by the formula E≈3μ, and a two-dimensional ultrasonic quantitative elastic image is obtained.

7. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 6, characterized in that: In step 2, the color coding technology converts the grayscale of the two-dimensional ultrasonic elastic image into a color image through a coding method, wherein the image part with a small elastic coefficient is encoded as red, indicating that the corresponding tissue is soft; the image part with a large elastic coefficient is encoded as blue, indicating that the corresponding tissue is hard; the image part with a medium elastic coefficient is encoded as green.

8. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 7, characterized in that: In step 3, the target graphics in the three-dimensional space are reconstructed based on the volume data, including: Step 3-1, construct the data structure: Determine the size of the reconstructed volume data according to the information of the two-dimensional ultrasound elastic image, the size scale includes the size of the volume data dimension, the coordinate origin information and the space interval between voxels, and determine the size of the reconstructed volume data by a bounding box-based technology; Step 3-2, reallocate sampling pixels: Traverse each pixel point on the two-dimensional elastic image, rearrange the position information of the pixel points through the transformation matrix, and map the pixel to the correct position of the three-dimensional volume data. The mapping method is forward mapping, including: in i =MX i (3) Among them, u i represents the corresponding voxel coordinates in the reconstructed volume data, M is the forward transformation matrix, X i is the pixel coordinate on the two-dimensional ultrasound quantitative elasticity image; Step 3-3, calculate the empty voxel value in the volume data: After the sampled pixels are redistributed as described in step 3-2, there are blank areas in the reconstructed volume, and the unknown voxel values ​​in the voxel grid are supplemented by interpolating the surrounding known pixel values.

9. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 8, characterized in that: The algorithm used in the volume data reconstruction method in step 3 is a pixel-based method, which is achieved by reallocating pixels and filling blank area voxels. The pixel reallocation stage adopts the nearest pixel assignment method, which assigns the nearest pixel value found after traversal to the voxel according to the forward mapping. If the voxel value is originally empty, it is directly assigned. If it is not empty, multiple pixel values ​​corresponding to the voxel are averaged and assigned. The traditional distance weighted algorithm is used to fill empty voxels in step 3-3, as follows: Among them, traverse the empty voxel points u that need to be filled j , with the voxel point as the center and r as the radius, determine a circular neighborhood N i , j∈N j is a set of known voxels in the neighborhood, according to the distance weighted rule ω j The weighted average of these known voxels is used to assign values ​​to the corresponding empty voxel points.

10. A three-dimensional ultrasonic shear wave elastic imaging method based on a linear array according to claim 9, characterized in that: In step 4, the method used by the three-dimensional visualization rendering technology is volume-based rendering, which uses rays to project the three-dimensional image and project it onto a plane for display; the voxels along each ray are weighted according to the following method: C out =C in [1-a(j)]+c j a j (5) Among them, C out is the value of the ray coming out of the jth voxel, C in is the value of the ray entering the jth voxel, parameter a controls the opacity, parameter c j It is a parameter adjusted based on the local gradient, controlling the brightness of the voxel; When a(j) = 0, the voxel is transparent and the ray is emitted through the jth voxel; when a(j) = 1, the parameter c j Determines whether the voxel is opaque or emissive.

Images