A clutter suppression method and system for fast real-time ultrasonic microvessel imaging

Abstract

The application discloses a clutter suppression method and system for fast real-time ultrasonic microvessel imaging, which reorganizes orthogonal demodulation data of a three-dimensional ultrasonic original image sequence obtained through multi-angle compound beam synthesis into a two-dimensional space-time matrix; according to prior information that tissue signals have low-rank space-time characteristics, a low-rank prior robust principal component analysis theoretical model is constructed to save operation steps for solving a low-rank matrix in a robust principal component analysis process; a convex optimization constraint solving function is constructed according to the theoretical model, and a fast alternating projection algorithm is adopted to obtain an iterative optimal solution of a sparse matrix of a blood flow signal subspace, thereby saving operation cost caused by global SVD in the iteration process. Through low-rank prior estimation and the fast alternating projection algorithm, the operation speed of ultrasonic microvessel imaging is improved while more vessel profile details are retained, problems existing in the prior art are solved, and an important means is provided for fast real-time clutter suppression clinical application of ultrasonic microvessel imaging.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of ultrasonic imaging, and particularly relates to a clutter suppression method and system for fast real-time ultrasonic microvessel imaging. BACKGROUND

[0002] Currently, the onset of many diseases in clinical practice is often closely related to the morphological changes or hemodynamic changes of microvessels. Non-invasively obtaining the dynamic characteristics of microvessels is an important challenge faced by medical imaging research. In recent years, the combination of ultrafast plane wave imaging technology and ultrasonic Doppler technology not only achieves a wide imaging field of view with a high imaging frame rate, but also provides rich spatiotemporal information, realizes the real-time detection and visualization of ultrasonic blood vessel structure for microvessels without invasion and contrast, and has important significance for the diagnosis, treatment and monitoring of clinical microvessel diseases. However, an important factor limiting the improvement of microvessel imaging quality is the clutter interference from tissue echoes or handheld probe artifacts and noise signals during imaging. The existence of large-amplitude clutter signals that drown the blood signals will reduce the sensitivity of microvessel detection. Therefore, an effective clutter suppression algorithm is an important means to improve the sensitivity of ultrasonic microvessel imaging.

[0003] The initial clutter suppression method adopts a traditional frequency domain filtering method, which is based on the distinction of the corresponding frequency domain ranges of the clutter signal, the blood flow signal and the noise signal in the slow time. However, this method is only effective in large blood vessels and areas with slow tissue movement. When the tissue moves, the tissue signal and the blood flow signal will partially overlap in the slow time frequency domain, greatly reducing the clutter suppression effect. The most widely used method for ultrasonic imaging clutter suppression is singular value decomposition (SVD) spatiotemporal filtering. The basic idea is to discard the tissue signal representing high correlation and the smallest component representing high-frequency noise signal, and to use the values of the eigenvectors corresponding to the remaining eigenvalues to reconstruct the blood signal spatiotemporal matrix of the blood vessel imaging. However, the SVD spatiotemporal filtering uses a hard threshold to separate the tissue signal and the blood flow signal subspace and the blood signal and the noise signal subspace. When the blood flow and the tissue signal overlap in the frequency domain, the selection deviation of the threshold will affect the clutter suppression effect, and the detection and filtering effect of the details and edges of the microvessels is limited. Robust principal component analysis (RPCA) is a process of clutter suppression by decomposing the original signal into a low-rank tissue component and a sparse blood flow component, which can significantly improve the sensitivity of blood vessel imaging detection. However, the traditional RPCA method relies on a large set of samples, which requires a large amount of computing power and storage, and has low operation efficiency. The existing clutter suppression techniques cannot balance the problems of fast real-time processing speed and rich blood flow detail information under a large set of samples, which affects the imaging quality of clinical ultrasonic microflow imaging. SUMMARY

[0004] The present application aims at solving the problem of low operation speed and inability to obtain detailed information in the prior art under the condition of large sample set for the clutter suppression link of ultrasonic microvessel detection, and provides a fast real-time ultrasonic microvessel imaging clutter suppression method and system.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0006] The fast real-time ultrasonic microvessel imaging clutter suppression method provided by the present application comprises the following steps:

[0007] Obtain a plurality of orthogonal demodulation original three-dimensional ultrasonic image sequences after beam synthesis of multi-angle ultrasonic signals, and reorganize the original three-dimensional ultrasonic image sequences into a two-dimensional space-time matrix; construct a low-rank priori robust principal component analysis theoretical model based on the two-dimensional space-time matrix;

[0008] Construct a minimum non-convex optimization function according to the low-rank priori robust principal component analysis theoretical model, and obtain a two-dimensional space-time matrix of ultrasonic microflow signals subjected to fast real-time clutter suppression when the minimum non-convex optimization function meets the convergence condition;

[0009] Reorganize the two-dimensional space-time matrix of ultrasonic microflow signals into a three-dimensional ultrasonic sequence, obtain the morphological structure of microvessels and the velocity of microflow through ultrasonic Doppler imaging, and obtain ultrasonic microflow imaging subjected to clutter suppression.

[0010] Preferably, the plurality of orthogonal demodulation original three-dimensional ultrasonic image sequences after beam synthesis of multi-angle ultrasonic signals are , and the obtained three-dimensional ultrasonic image sequence is reorganized into a two-dimensional space-time matrix by Casorati transformation.

[0011] Wherein, represents a complex domain, represents the number of axial samples in the spatial dimension, and is perpendicular to the sensor array; represents the number of transverse samples in the spatial dimension, and is parallel to the sensor array; represents the number of samples in the time dimension. represents the number of samples in the spatial dimension of the reorganized two-dimensional space-time matrix .

[0012] Preferably, the two-dimensional space-time matrix is decomposed into an organization component with low-rank characteristics and a blood flow component with sparse characteristics according to different space-time characteristics.

[0013] wherein the tissue component the rank According to the different imaging area in advance, the rank estimate is the low rank prior information.

[0014] Preferably, according to the difference of the spatio-temporal coherence of the tissue signal and the blood flow signal in the two-dimensional spatio-temporal matrix, the two-dimensional spatio-temporal matrix is decomposed by low rank and sparsity, and a low rank prior robust principal component analysis theoretical model is established by using low rank prior information constraint as follows:

[0015]

[0016] wherein, respectively represent the low rank tissue matrix and the sparse blood flow matrix, represents the nuclear norm, represents the -norm, is a weight coefficient, , represents the rank of the matrix, represents the prior rank estimate value of the low rank tissue matrix , the number of samples in the time dimension, represents the number of samples in the spatial dimension of the recombined two-dimensional spatio-temporal matrix .

[0017] Preferably, the minimum non-convex optimization function is constructed as follows:

[0018]

[0019] wherein, represents the -norm; is a two-dimensional spatio-temporal matrix, respectively represent the low rank tissue matrix and the sparse blood flow matrix, represents the prior rank estimate value of the low rank tissue matrix .

[0020] The minimum non-convex optimization function is iteratively solved by using a fast alternating projection algorithm, and when the convergence condition is met, the two-dimensional spatio-temporal matrix of the ultrasonic micro blood flow signal is obtained.

[0021] Preferably, in the iteration process, represents the variable iteration result of the first time, represents the variable iteration result of the first time;

[0022] is updated Low-rank organization subspace at the next iteration :

[0023]

[0024] where, is the matrix projection operator, is the projected subspace; is the truncated SVD operator; to avoid introducing the global SVD calculation with high computational complexity, the QR decomposition is introduced to simplifies to:

[0025]

[0026] where, , and are the spatial vector matrix, diagonal matrix and time vector matrix after singular value decomposition, respectively; is the orthogonal triangular (QR) decomposition result of , is the orthogonal triangular (QR) decomposition result of , is the orthogonal triangular (QR) decomposition result of , is the low-rank intermediate matrix , is the transpose of matrix

[0027] update Sparse blood flow subspace at the next iteration :

[0028]

[0029] where, is the hard threshold shrinkage operator, is the hard threshold;

[0030]

[0031] where, is the adjustment parameter, taking values , is the convergence speed parameter, ranging from 0 to 1, and represent the first singular value and the th singular value, respectively;

[0032] The relative error of iteration is less than the error limit , the iteration convergence condition is met, and the iteration is terminated to obtain the final two-dimensional spatio-temporal matrix of ultrasonic micro blood flow signals ; wherein, error limit .

[0033] Preferably, the iterative optimal solution ultrasonic micro blood flow signal two-dimensional space-time matrix Reorganized into a three-dimensional space-time sequence by Casorati inverse transformation , the clutter suppressed ultrasonic vascular morphology map is obtained by ultrasonic power Doppler, and ultrasonic imaging clutter suppression is realized.

[0034] The application provides a fast real-time ultrasonic micro blood vessel imaging clutter suppression system, which comprises:

[0035] A model construction module is configured to obtain a plurality of frames of orthogonal demodulated original three-dimensional ultrasonic image sequences after beam synthesis of multi-angle ultrasonic signals, reorganize the original three-dimensional ultrasonic image sequences into a two-dimensional space-time matrix, and construct a low-rank prior robust principal component analysis theoretical model based on the two-dimensional space-time matrix.

[0036] A fast real-time clutter suppression module is configured to construct a minimum non-convex optimization function according to the low-rank prior robust principal component analysis theoretical model, and obtain a two-dimensional space-time matrix of ultrasonic micro blood flow signals subjected to fast real-time clutter suppression when the minimum non-convex optimization function meets a convergence condition.

[0037] An image acquisition module is configured to reorganize the two-dimensional space-time matrix of ultrasonic micro blood flow signals into a three-dimensional ultrasonic sequence, obtain the morphology structure of micro blood vessels and the velocity of micro blood flow through ultrasonic Doppler imaging, and acquire ultrasonic micro blood flow imaging subjected to clutter suppression.

[0038] A computer device comprises a memory and a processor, the memory stores a computer program, and the processor realizes the steps of the fast real-time ultrasonic micro blood vessel imaging clutter suppression method when executing the computer program.

[0039] A computer readable storage medium stores a computer program, and the computer program realizes the steps of the fast real-time ultrasonic micro blood vessel imaging clutter suppression method when executed by a processor.

[0040] Compared with the prior art, the application has the following beneficial effects:

[0041] This invention proposes a rapid real-time ultrasound microvascular imaging clutter suppression method. The method first reconstructs the original three-dimensional ultrasound sequence into a two-dimensional spatiotemporal matrix. Based on the low-rank characteristics of tissue signals and the sparsity of blood flow signals, the rank of the tissue signal is predicted, and this prediction is used as prior information to construct a low-rank prior robust principal component analysis (LPA) theoretical model, saving the computational steps of solving the low-rank matrix during LPA. Based on the low-rank prior robust LPA theoretical model, a corresponding minimum non-convex optimization function is constructed. Solving this function yields the clutter-suppressed two-dimensional spatiotemporal matrix of the blood flow signal. This matrix is ​​then converted back into a three-dimensional ultrasound sequence to obtain the morphological features and velocity information of the micro-blood flow. Specifically, in constructing the low-rank prior robust LPA theoretical model, this invention saves the low-rank matrix solution process during iteration by using low-rank prior estimation. By constructing the minimum non-convex optimization function and employing a fast alternating projection algorithm for iterative solution, the computational complexity is greatly reduced. This method can significantly improve the computational speed of ultrasound microvascular imaging in the clutter suppression stage, enabling rapid and real-time acquisition of ultrasound microvascular information, solving the problems existing in the current technology, and providing an important means for clinical ultrasound microvascular imaging.

[0042] This invention proposes a rapid real-time ultrasound microvascular imaging clutter suppression system. By dividing the system into a model building module, a rapid real-time clutter suppression module, and an image acquisition module, it achieves real-time ultrasound imaging clutter suppression. The modular approach ensures that each module is independent, facilitating unified management of all modules. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a flowchart of the rapid real-time ultrasound microvascular imaging clutter suppression method of the present invention.

[0045] Figure 2 This is a diagram showing the clutter suppression phantom result of a high-sensitivity, fast, real-time ultrasound vascular imaging system in a specific embodiment of the present invention.

[0046] Figure 3 This is a diagram of the rapid real-time ultrasound microvascular imaging clutter suppression system of the present invention. Detailed Implementation

[0047] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings. Obviously, the described embodiments are only some but not all of the embodiments of the present application. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0048] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative work fall within the scope of protection of the present application.

[0049] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0050] In the description of the embodiments of the present application, it should be noted that if the terms "upper", "lower", "horizontal", "inner" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship when the product of the present application is used, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the indicated device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second" and the like are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0051] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly inclined. For example, "horizontal" only means that its direction is relatively more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly inclined.

[0052] In the description of the embodiments of the present application, it should also be noted that unless otherwise explicitly specified and limited, if the terms "arrangement", "installation", "connection", "connection" appear, they should be understood in a broad sense, for example, they can be fixedly connected, or detachably connected, or integrally connected; can be mechanically connected, or electrically connected; can be directly connected, or indirectly connected through an intermediate medium, or the communication inside two elements. For those of ordinary skill in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0053] The present application will be described in further detail below with reference to the accompanying drawings:

[0054] The application provides a clutter suppression method for fast real-time ultrasonic microvessel imaging. Figure 1 As shown in the formula, the method comprises the following steps:

[0055] S1, obtaining a plurality of multi-angle ultrasonic signal beam synthesis orthogonal demodulation original three-dimensional ultrasonic image sequences, recombining the original three-dimensional ultrasonic image sequences into a two-dimensional space-time matrix, and constructing a low-rank priori robust principal component analysis theoretical model based on the two-dimensional space-time matrix;

[0056] The plurality of multi-angle ultrasonic signal beam synthesis orthogonal demodulation original three-dimensional ultrasonic image sequences are The obtained three-dimensional ultrasonic image sequences are recombined into a two-dimensional space-time matrix by Casorati transformation.

[0057] Wherein, represents a complex domain, represents an axial sample number in a spatial dimension, and a vertical sensor array; represents a transverse sample number in a spatial dimension, and a parallel sensor array; represents a sample number in a time dimension. represents a sample number in a spatial dimension of the recombined two-dimensional space-time matrix .

[0058] The two-dimensional space-time matrix is decomposed into an organization component with a low-rank characteristic and a blood flow component with a sparse characteristic according to different space-time characteristics.

[0059] Wherein, the rank of the organization component is estimated in advance according to different imaging regions, and the rank estimation value is low-rank priori information; in a phantom experiment, the range of the low-rank priori estimation is 1-5; in a living body experiment, the range of the low-rank priori estimation of a region with small tissue motion, for example, a brain and a thyroid, is 5-10; and the range of the low-rank priori estimation of a region with large tissue motion, for example, a heart and a kidney, is 10-20.

[0060] According to the space-time coherence difference between the tissue signal and the blood flow signal in the two-dimensional space-time matrix, the two-dimensional space-time matrix is subjected to low-rank sparse decomposition, and a low-rank priori robust principal component analysis theoretical model is established by using low-rank priori information constraint as follows:

[0061] ​

[0062] where, respectively represent the low-rank tissue matrix and the sparse blood flow matrix, denotes the nuclear norm, denotes the -norm, is a weight coefficient, , denotes the rank of a matrix, denotes the prior rank estimation of the low-rank tissue matrix , denotes the number of samples in the time dimension, denotes the number of samples in the spatial dimension of the reorganized two-dimensional space-time matrix .

[0063] S2, constructing a minimum non-convex optimization function according to the low-rank prior robust principal component analysis theoretical model, when the minimum non-convex optimization function meets the convergence condition, obtaining the two-dimensional space-time matrix of the ultrasonic micro blood flow signal after fast real-time clutter suppression;

[0064] The constructed minimum non-convex optimization function is as follows:

[0065]

[0066] where, denotes the -norm;

[0067] The fast alternating projection algorithm is used to iteratively solve the minimum non-convex optimization function, and when the convergence condition is met, the two-dimensional space-time matrix of the ultrasonic micro blood flow signal is obtained;

[0068] In the iteration process, denotes the variable iteration result of the m-th time, denotes the variable iteration result of the m-th time; update the low-rank tissue subspace of the m-th iteration:

[0069]

[0070]

[0071] where, represents a matrix projection operator, is a projected subspace; denotes a truncated SVD operator. In order to avoid introducing full-size SVD calculation with high computational complexity, QR decomposition is introduced to simplify the solution of to:

[0072] ​​​

[0073] in, , and They are respectively The spatial vector matrix, diagonal matrix, and time vector matrix after singular value decomposition; yes The orthogonal triangular (QR) decomposition results, yes The orthogonal triangular (QR) decomposition result; defining the low-rank intermediate matrix. ,but yes The spatial vector matrix, diagonal matrix, and time vector matrix after singular value decomposition; Represents the transpose of a matrix;

[0074] renew Sparse blood flow subspace at the next iteration :

[0075]

[0076] in, For hard threshold shrinkage operator, Hard threshold;

[0077]

[0078] in, To adjust the parameters, take the following values: , This is the convergence speed parameter, ranging from 0 to 1, with a default value of 0.5. and Representing the first singular value and the second singular value respectively. One singular value;

[0079] Iteration relative error Less than the error limit When the iterative convergence condition is met, the iteration is terminated to obtain the final two-dimensional spatiotemporal matrix of the ultrasound micro-blood flow signal. Among them, the error limit .

[0080] S3. The two-dimensional spatiotemporal matrix of the ultrasound micro-blood flow signal is reconstructed into a three-dimensional ultrasound sequence. The morphology and structure of micro-vessels and the velocity of micro-blood flow are obtained through ultrasound Doppler imaging, and ultrasound micro-blood flow images after clutter suppression are acquired.

[0081] The iterative optimal solution is obtained from the two-dimensional spatiotemporal matrix of the ultrasound micro-blood flow signal. Reconstructed into a three-dimensional spatiotemporal sequence by Casarati inverse transform , and the ultrasonic blood vessel shape image after clutter suppression is obtained by ultrasonic power Doppler, so as to realize ultrasonic imaging clutter suppression.

[0082] The detailed steps are as follows:

[0083] Step 1, using an ultrasonic instrument to collect ultrasonic signals in an imaging area, performing multi-frame multi-angle beam synthesis, and obtaining a three-dimensional ultrasonic image sequence by Casorati transformation , wherein represents the number of axial samples in the spatial dimension (vertical sensor array), represents the number of transverse samples in the spatial dimension (parallel sensor array), represents the number of samples in the time dimension (time frame number), represents the number of samples in the spatial dimension of the reorganized two-dimensional space-time matrix .

[0084] Step 2, according to the difference in space-time coherence between tissue signals and blood flow signals, the tissue component usually has a high-amplitude low-rank characteristic, and the blood flow component presents a low-amplitude sparse characteristic, the two-dimensional space-time matrix is subjected to low-rank sparse decomposition, and a low-rank priori robust principal component analysis theoretical model is established by using low-rank priori information constraint:

[0085]

[0086] , wherein respectively represent a low-rank tissue matrix and a sparse blood flow matrix, represents a nuclear norm, represents an -norm, is a weight coefficient, , represents the rank of a matrix, represents the priori rank estimation value of the low-rank tissue matrix , the number of samples in the time dimension, represents the number of samples in the spatial dimension of the reorganized two-dimensional space-time matrix .

[0087] Step 3, the process of solving the model is a low-rank sparse matrix optimal solution solving problem based on low-rank priori constraint: in order to reduce the low-rank tissue matrix The calculation complexity of the projection is based on an alternating projection algorithm for solving a low-rank prior robust principal component analysis theoretical model by alternately projecting a matrix residual into a low-rank matrix set and a sparse matrix set, a minimum non-convex optimization function is constructed according to the low-rank prior robust principal component analysis theoretical model, and when the minimum non-convex optimization function meets a convergence condition, a two-dimensional space-time matrix of the ultrasonic micro blood flow signal subjected to rapid real-time clutter suppression is obtained;

[0088] The minimum non-convex optimization function is iteratively solved by using the fast alternating projection algorithm, and when the convergence condition is met, the two-dimensional space-time matrix of the ultrasonic micro blood flow signal is obtained.

[0089] The constructed minimum non-convex optimization function is as follows:

[0090]

[0091] Wherein, represents a norm;

[0092] In the iteration process, represents the variable in the first iteration result, represents the variable in the first iteration result;

[0093] The low-rank tissue subspace in the first iteration is updated as :

[0094]

[0095] Wherein, represents a matrix projection operator, is a projected subspace; represents a truncated SVD operator. In order to avoid introducing a global SVD calculation with high calculation complexity, a QR decomposition is introduced to simplify the solution of :

[0096]

[0097] Wherein, , and are respectively a space vector matrix, a diagonal matrix and a time vector matrix after singular value decomposition; is an orthogonal triangular (QR) decomposition result of , is an orthogonal triangular (QR) decomposition result of ; define a low-rank intermediate matrix , then is The spatial vector matrix, diagonal matrix, and time vector matrix after singular value decomposition; Represents the transpose of a matrix;

[0098] renew Sparse blood flow subspace at the next iteration :

[0099]

[0100] in, For hard threshold shrinkage operator, Hard threshold;

[0101]

[0102] in, To adjust the parameters, take the following values: , This is the convergence speed parameter, ranging from 0 to 1, with a default value of 0.5. and Representing the first singular value and the second singular value respectively. One singular value;

[0103] Iteration relative error Less than the error limit When the iterative convergence condition is met, the iteration is terminated to obtain the final two-dimensional spatiotemporal matrix of the ultrasound micro-blood flow signal. Among them, the error limit .

[0104] Step 4: Reconstruct the two-dimensional spatiotemporal matrix of the ultrasound micro-blood flow signal into a three-dimensional ultrasound sequence, and obtain the morphology and structure of micro-vessels and the velocity of micro-blood flow through ultrasound Doppler imaging to achieve ultrasound micro-blood flow imaging.

[0105] The iterative optimal solution is recombined into a three-dimensional spatiotemporal sequence through the Cascoati inverse transform, and the ultrasound vascular morphology map after clutter suppression is obtained by ultrasound power Doppler, thus realizing ultrasound imaging clutter suppression.

[0106] The following analysis will be based on specific cases:

[0107] To investigate the effect of ultrasound imaging clutter suppression, the high-sensitivity, fast, real-time ultrasound vascular imaging clutter suppression method proposed in this invention was used to process simulated blood flow Doppler imaging results of a vascular phantom at a flow rate of 1 mm / s. The phantom was a wallless phantom of a 0.8 mm blood vessel, prepared from a 10% gelatin solution and 2% agar. Five angles with a center frequency of 7 MHz were transmitted using a programmable ultrasound device. A 1-cycle multi-angle ultrasonic plane wave signal with a pulse repetition frequency of 2.5 kHz was used to synthesize a 250-frame ultrasonic image sequence using Verasonics multi-angle plane wave beamforming.

[0108] Figure 2 This is a vascular phantom imaging result with clutter suppression for high-sensitivity, fast, real-time ultrasound vascular imaging. Figure 2 Image 'a' is the original ultrasound image. It can be seen that the presence of clutter makes it impossible to accurately detect tiny blood vessels. Figure 2 b is the ultrafast power Doppler image after clutter suppression using the clutter suppression method described in this invention. It significantly improves the contrast and average level of blood flow signals, and greatly reduces the computational complexity compared to the traditional RPCA method. It can achieve high-sensitivity, fast, and real-time clutter suppression in ultrasound vascular imaging, thereby improving the quality of ultrasound microvascular imaging.

[0109] This invention proposes a high-sensitivity, fast, real-time ultrasound vascular imaging clutter suppression system, such as... Figure 3 As shown, it includes a model building module, a fast real-time clutter suppression module, and an image acquisition module;

[0110] The model building module is used to obtain the orthogonally demodulated original three-dimensional ultrasound image sequence after beamforming of multi-frame, multi-angle ultrasound signals, and to reassemble the original three-dimensional ultrasound image sequence into a two-dimensional spatiotemporal matrix; a low-rank prior robust principal component analysis theoretical model is constructed based on the two-dimensional spatiotemporal matrix.

[0111] The fast real-time clutter suppression module is used to construct a minimum non-convex optimization function based on the low-rank prior robust principal component analysis theoretical model. When the minimum non-convex optimization function satisfies the convergence condition, the two-dimensional spatiotemporal matrix of the ultrasound micro-blood flow signal after fast real-time clutter suppression is obtained.

[0112] The image acquisition module is used to reconstruct the two-dimensional spatiotemporal matrix of the ultrasound micro-blood flow signal into a three-dimensional ultrasound sequence, obtain the morphology and structure of micro-vessels and the velocity of micro-blood flow through ultrasound Doppler imaging, and acquire ultrasound micro-blood flow images after clutter suppression.

[0113] The terminal device provided in this embodiment of the invention includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the various method embodiments described above. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the various device embodiments described above.

[0114] The computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention.

[0115] The terminal device can be a desktop computer, a notebook computer, a palm computer, a cloud server, and the like. The terminal device can include, but is not limited to, a processor and a memory.

[0116] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like.

[0117] The memory can be used to store the computer programs and / or modules, and the processor realizes various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory, and calling data stored in the memory.

[0118] The modules / units integrated in the terminal device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the processor executes the computer program, the steps of the above-mentioned various method embodiments can be realized. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0119] The application provides a fast real-time ultrasonic microvessel imaging clutter suppression method. The method is based on a low-rank prior robust principal component analysis theoretical model and uses a fast alternating projection algorithm to solve a constraint optimization problem, so that an iterative optimal solution of a sparse matrix corresponding to a blood flow signal subspace is converted into a three-dimensional ultrasonic blood flow signal sequence, and a clutter suppression process is realized. The low-rank prior robust principal component analysis-based high-sensitivity fast real-time ultrasonic vascular imaging clutter suppression method improves the sensitivity and signal-to-noise ratio of ultrasonic microvessel imaging, accelerates the clutter suppression iterative calculation speed, and provides an important means for fast real-time improvement of the sensitivity of ultrasonic microvessel imaging.

[0120] The above merely illustrates the preferred embodiments of the present application, and is not intended to limit the present application. The present application can be variously changed and modified by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A clutter suppression method for fast real-time ultrasonic microvascular imaging, characterized in that, The method comprises the following steps: obtaining a sequence of original three-dimensional ultrasound images after orthogonal demodulation and beamforming of multi-frame multi-angle ultrasound signals, reorganizing the sequence of original three-dimensional ultrasound images into a two-dimensional space-time matrix, and constructing a low-rank priori robust principal component analysis theoretical model based on the two-dimensional space-time matrix; constructing a minimum non-convex optimization function according to the low-rank priori robust principal component analysis theoretical model, and obtaining a two-dimensional space-time matrix of the ultrasound micro blood flow signal after fast real-time clutter suppression when the minimum non-convex optimization function meets a convergence condition; reorganizing the two-dimensional space-time matrix of the ultrasound micro blood flow signal into a three-dimensional ultrasound sequence, obtaining the morphology and structure of the micro blood vessels and the speed of the micro blood flow through ultrasound Doppler imaging, and obtaining ultrasound micro blood flow imaging after clutter suppression; Two-dimensional space-time matrix According to the different space-time characteristics, decompose into the organization component with low rank characteristics And the blood flow component with sparse characteristics ; wherein, the rank of the organization component According to the different imaging area, make an estimate in advance, the rank estimate value Is the low rank prior information; according to the space-time coherence difference of the organization signal and the blood flow signal in the two-dimensional space-time matrix Carry out low rank sparse decomposition, and use the low rank prior information constraint to establish the low rank prior robust principal component analysis theory model as follows:​ wherein, respectively denote the low-rank tissue matrix and the sparse blood flow matrix, denotes the nuclear norm, denotes the p-norm, is a weight coefficient, , denotes the rank of a matrix, denotes the prior rank estimate of the low-rank tissue matrix , denotes the number of samples in the temporal dimension, denotes the number of samples in the spatial dimension of the reorganized two-dimensional spatio-temporal matrix .

2. The clutter suppression method for fast real-time ultrasonic microvascular imaging according to claim 1, characterized in that, The orthogonal demodulation original three-dimensional ultrasound image sequence after the beam synthesis of the multi-frame multi-angle ultrasound signals is The obtained three-dimensional ultrasound image sequence is reorganized into a two-dimensional space-time matrix ; wherein represents the complex domain, represents the number of axial samples in the spatial dimension, perpendicular sensor array; represents the number of lateral samples in the spatial dimension, parallel sensor array; represents the number of samples in the temporal dimension; represents the number of samples in the spatial dimension of the reorganized two-dimensional space-time matrix .

3. The clutter suppression method for fast real-time ultrasonic microvascular imaging according to claim 1, wherein, the constructed minimum non-convex optimization function is as follows: wherein, denotes - a norm; is a two-dimensional space-time matrix, denote a low-rank tissue matrix and a sparse blood flow matrix, respectively, denotes a low-rank tissue matrix a priori rank estimate value; iteratively solving the minimum non-convex optimization function by using a fast alternating projection algorithm, and obtaining the two-dimensional space-time matrix of the ultrasound micro blood flow signal when a convergence condition is met.

4. The clutter suppression method for fast real-time ultrasonic microvascular imaging according to claim 3, wherein, during the iteration, denotes the variable of the iteration, denotes the variable of the iteration; updating low rank organization subspace at sub-iterations : where denotes the matrix projection operator, is the projected subspace; denotes the truncated SVD operator; to avoid introducing a computationally expensive global SVD computation, a QR decomposition is introduced to the solution simplifies to: in, , and They are respectively The spatial vector matrix, diagonal matrix, and time vector matrix after singular value decomposition; yes The orthogonal triangular QR decomposition results, yes The orthogonal triangular QR decomposition results; defining the low-rank intermediate matrix. ,but yes The spatial vector matrix, diagonal matrix, and time vector matrix after singular value decomposition; Represents the transpose of a matrix; updating sparse blood flow subspace at the sub-iteration : wherein, is a hard thresholding operator, is a hard thresholding. wherein, is a tuning parameter, taking values , is a convergence speed parameter, ranging from 0 to 1, and denote the 1st and the th singular values, respectively. Iterative relative error Less than error limit When the iteration converges, the iteration is terminated to obtain a final ultrasonic micro blood flow signal two-dimensional space-time matrix ; wherein the error limit .

5. The clutter suppression method for fast real-time ultrasonic microvascular imaging according to claim 1, wherein, Two-dimensional space-time matrix of ultrasonic micro blood flow signal iteration optimal solution Reorganization into three-dimensional space-time sequence by Casorati inverse transformation The ultrasonic blood vessel morphology diagram after clutter suppression is obtained through ultrasonic power Doppler, and ultrasonic imaging clutter suppression is realized.

6. A clutter rejection system for fast real-time ultrasonic microvascular imaging, characterized in that, The method comprises the following steps: a model construction module configured to obtain a sequence of original three-dimensional ultrasound images after orthogonal demodulation and beamforming of multi-frame multi-angle ultrasound signals, reorganize the sequence of original three-dimensional ultrasound images into a two-dimensional space-time matrix, and construct a low-rank priori robust principal component analysis theoretical model based on the two-dimensional space-time matrix; a fast real-time clutter suppression module configured to construct a minimum non-convex optimization function according to the low-rank priori robust principal component analysis theoretical model, and obtain a two-dimensional space-time matrix of the ultrasound micro blood flow signal after fast real-time clutter suppression when the minimum non-convex optimization function meets a convergence condition; an image acquisition module configured to reorganize the two-dimensional space-time matrix of the ultrasound micro blood flow signal into a three-dimensional ultrasound sequence, obtain the morphology and structure of the micro blood vessels and the speed of the micro blood flow through ultrasound Doppler imaging, and obtain ultrasound micro blood flow imaging after clutter suppression. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the fast real-time ultrasound micro blood vessel imaging clutter suppression method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, wherein the computer program comprises the following steps of: The computer program is executed by the processor to implement the steps of the fast real-time ultrasound micro blood vessel imaging clutter suppression method according to any one of claims 1 to 5.

Images