Cluster-based magnetic resonance diffusion ordered spectroscopy processing method and apparatus, and readable medium

Abstract

The application discloses a kind of based on clustering's magnetic resonance diffusion ordering spectrum processing method, device and readable medium, by obtaining the attenuation spectrum corresponding to several gradient coding magnetic resonance diffusion ordering spectrum, selecting the signal point of each spectral peak on the chemical shift dimension in attenuation spectrum, and constructing the feature vector corresponding to each signal point, length is the number of gradients;According to the feature vector of all signal points of each spectral peak, classification is carried out by clustering algorithm, and the classification result of the signal point of each spectral peak is obtained;According to the classification result of the signal point of each spectral peak and the spectral intensity corresponding to each signal point, statistical analysis is carried out, and the class of each spectral peak is determined, according to the class, the magnetic resonance diffusion ordering spectrum is separated into components, and the separated spectrum corresponding to different components is obtained.The method is simple in operation, widely applicable, excellent in effect, suitable for a small amount of gradient, can effectively shorten time, good in robustness, and small error.

Description

Technical Field

[0001] This invention relates to the field of magnetic resonance diffusion sorting spectrum processing, and specifically to a cluster-based magnetic resonance diffusion sorting spectrum processing method, apparatus, and readable medium. Background Technology

[0002] Magnetic resonance diffusion ordering spectroscopy relies on the diffusion properties of molecules to perform a virtual separation of molecular components in a mixture, meaning that the spectral peaks are separated but the sample itself is not. Magnetic resonance diffusion ordering spectroscopy is widely used in industrial detection, biochemistry, and clinical medicine for the analysis and identification of components in mixtures. Existing diffusion ordering spectroscopy processing methods rely on quantitatively calculating the diffusion coefficient from the gradient decay signal to separate molecular components. However, quantitatively calculating the diffusion coefficient requires a long gradient decay signal, meaning that the signal needs to be repeatedly acquired using different gradient codes in the experiment, resulting in long experimental times. Especially in high-dimensional magnetic resonance diffusion ordering spectroscopy experiments, multiple gradients lead to a sharp increase in time costs. Furthermore, since calculating the decay constant from the exponential decay signal is mathematically ill-posed, existing diffusion ordering spectroscopy processing methods often lack robustness, leading to large errors in component separation results.

[0003] Therefore, the industry urgently needs a fast and robust magnetic resonance diffusion ordering spectroscopy processing method based on a small number of gradients. This method can further advance the application of magnetic resonance diffusion ordering spectroscopy in scientific research and industry. Summary of the Invention

[0004] In view of the aforementioned problems with existing magnetic resonance diffusion sorting spectroscopy processing methods, such as the need for multiple dimensions, long processing times, lack of robustness, and large errors, the purpose of this application is to propose a clustering-based magnetic resonance diffusion sorting spectroscopy processing method, apparatus, and readable medium to solve the technical problems mentioned in the background section.

[0005] In a first aspect, the present invention provides a cluster-based magnetic resonance diffusion sorting spectrum processing method, comprising the following steps:

[0006] S1, obtain the attenuation spectrum corresponding to the magnetic resonance diffusion sorting spectrum of several gradients;

[0007] S2, select the signal point of each spectral peak in the chemical shift dimension in the decay spectrum, and construct a feature vector corresponding to each signal point with a length equal to the number of gradients;

[0008] S3. Based on the feature vectors of all signal points of each spectral peak, a clustering algorithm is used to classify the signal points of each spectral peak to obtain the classification result of each spectral peak.

[0009] S4. Based on the classification results of the signal points of each spectral peak and the spectral intensity corresponding to each signal point, statistical analysis is performed to determine the category of each spectral peak. Based on the category, the magnetic resonance diffusion sorting spectrum is separated into components to obtain the separated spectrum corresponding to different components.

[0010] Preferably, in step S1, the attenuation spectrum is encoded using j gradients, the attenuation signal consists of i components, the chemical shift length is p, and the data matrix Y of the attenuation spectrum is represented as follows:

[0011]

[0012] in, Let D be the exponential decay matrix of the components, where the k-th column of D represents the decay signal of the k-th component; D k b is the attenuation coefficient of the k-th component; k Let b be the k-th value, and b is proportional to the square of the gradient; S is the spectral matrix of the components, and the k-th row of S is the spectral signal of the k-th component.

[0013] Preferably, step S2 specifically includes:

[0014] The chemical shift spectrum of the first gradient encoding is normalized, and data points with spectral intensities greater than a threshold are used as signal points.

[0015] Obtain the index value of the signal point in the attenuation spectrum, extract the corresponding column vector in the data matrix Y according to the index value, and normalize the column vector by the maximum value and take the natural logarithm to form the feature vector corresponding to each signal point.

[0016] Preferred clustering algorithms include CFSFDP clustering algorithm, K-Means clustering algorithm, K-Medoids clustering algorithm, agglomerative hierarchical clustering algorithm, and split hierarchical clustering algorithm.

[0017] Preferably, in step S3, the CFSFDP clustering algorithm is used to classify all feature vectors according to their similarity to obtain the category of each signal point, where each category corresponds to a component.

[0018] Preferably, step S4 specifically includes:

[0019] The classification weight vector of each spectral peak is obtained based on the category and spectral intensity of each signal point;

[0020] Each spectral peak is assigned a class based on the classification weight vector, thus obtaining the class of each spectral peak;

[0021] The magnetic resonance diffusion sorting spectrum is separated into components based on the category of each spectral peak to obtain the separated spectrum.

[0022] Preferably, the elements in the classification weight vector are the sum of the spectral intensities of all signal points under each category, and the category corresponding to the largest value among the elements of the classification weight vector is the genus of each spectral peak.

[0023] Secondly, the present invention provides a cluster-based magnetic resonance diffusion sorting spectrum processing device, comprising:

[0024] The encoding module is configured to acquire the attenuation spectrum corresponding to the magnetic resonance diffusion sorting spectrum with several gradient encodings.

[0025] The feature vector extraction module is configured to select the signal point of each spectral peak in the chemical shift dimension of the decay spectrum and construct a feature vector corresponding to each signal point with a length equal to the number of gradients.

[0026] The clustering module is configured to classify the signal points of each spectral peak using a clustering algorithm based on the feature vectors of all signal points of each spectral peak, and obtain the classification result of the signal points of each spectral peak.

[0027] The separation module is configured to perform statistical analysis based on the classification results of the signal points of each spectral peak and the spectral intensity corresponding to each signal point to determine the category of each spectral peak, and then perform component separation of the magnetic resonance diffusion sorting spectrum according to the category to obtain the separation spectrum corresponding to different components.

[0028] Thirdly, the present invention provides an electronic device including one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0029] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any of the implementations of the first aspect.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] (1) The clustering-based magnetic resonance diffusion sorting spectrum processing method proposed in this invention can realize the separation processing of various magnetic resonance diffusion sorting spectrum components based on a small number of gradients. Signal points in the chemical shift dimension are selected on the decay spectrum, and the signal values ​​corresponding to each gradient encoding are used as feature values ​​to construct feature vectors. Based on the feature vectors, a clustering algorithm is used to classify the signal points. Statistical analysis is performed based on the classification results of the signal points contained in the spectral peaks to finally determine the class affiliation of each spectral peak.

[0032] (2) The cluster-based magnetic resonance diffusion sorting spectrum processing method proposed in this invention performs qualitative component separation based on the similarity of gradient signals. Only a small number of gradients (greater than or equal to 2) are needed to achieve component separation, thereby reducing experimental time.

[0033] (3) The cluster-based magnetic resonance diffusion sorting spectrum processing method proposed in this invention is simple to operate, widely applicable, and has excellent results. It is suitable for a small number of gradients, can effectively shorten the time, has good robustness, and has small error in the separation results. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is an exemplary device architecture diagram in which an embodiment of this application can be applied;

[0036] Figure 2 This is a schematic flowchart of a cluster-based magnetic resonance diffusion sorting spectrum processing method according to an embodiment of this application.

[0037] Figure 3 This is a schematic diagram illustrating the processing procedure of the cluster-based magnetic resonance diffusion sorting spectrum processing method according to an embodiment of this application.

[0038] Figure 4 This is a schematic diagram of a cluster-based magnetic resonance diffusion sorting spectrum processing device according to an embodiment of this application;

[0039] Figure 5 This is a schematic diagram of the structure of a computer device suitable for implementing the electronic device of the present application. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0041] Figure 1 An exemplary device architecture 100 is shown that can be applied to the cluster-based magnetic resonance diffusion sorting spectrum processing method or the cluster-based magnetic resonance diffusion sorting spectrum processing apparatus of the present application embodiments.

[0042] likeFigure 1 As shown, the device architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the terminal devices 101, 102, and 103 and the server 105. The network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0043] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various applications, such as data processing applications and file processing applications, can be installed on terminal devices 101, 102, and 103.

[0044] Terminal devices 101, 102, and 103 can be either hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices, including but not limited to smartphones, tablets, laptops, and desktop computers. When terminal devices 101, 102, and 103 are software, they can be installed in the electronic devices listed above. They can be implemented as multiple software programs or software modules (e.g., software programs or software modules used to provide distributed services) or as a single software program or software module. No specific limitations are imposed here.

[0045] Server 105 can be a server that provides various services, such as a background data processing server that processes files or data uploaded by terminal devices 101, 102, and 103. The background data processing server can process the acquired files or data and generate processing results.

[0046] It should be noted that the cluster-based magnetic resonance diffusion sorting spectrum processing method provided in this application embodiment can be executed by server 105 or by terminal devices 101, 102, and 103. Correspondingly, the cluster-based magnetic resonance diffusion sorting spectrum processing device can be set in server 105 or in terminal devices 101, 102, and 103.

[0047] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Any number of terminal devices, networks, and servers can be included depending on implementation needs. If the data being processed does not need to be retrieved remotely, the above architecture may not include a network, requiring only servers or terminal devices.

[0048] Figure 2 The present application illustrates a cluster-based magnetic resonance diffusion sorting spectrum processing method, comprising the following steps:

[0049] S1, obtain the attenuation spectrum corresponding to several gradient-encoded magnetic resonance diffusion sorting spectra.

[0050] In a specific embodiment, the attenuation spectrum in step S1 is encoded using j gradients, the attenuation signal consists of i components, the chemical shift length is p, and the data matrix Y of the attenuation spectrum is represented as follows:

[0051]

[0052] in, Let D be the exponential decay matrix of the components, where the k-th column of D represents the decay signal of the k-th component; D k b is the attenuation coefficient of the k-th component; k Let b be the k-th value, and b is proportional to the square of the gradient; S is the spectral matrix of the components, and the k-th row of S is the spectral signal of the k-th component.

[0053] Specifically, let's take one of the cases as an example, for reference. Figure 3 (a) In the embodiments of this application, the attenuation spectrum is encoded using three gradients, with thirty spectral peaks at the chemical shift. The spectral peak signal consists of three components, and the data matrix is ​​as follows:

[0054]

[0055] in, This is the exponential decay matrix of the components, where the k-th column of D represents the decay signal of the k-th component; D k Let D1 be the attenuation coefficient (or diffusion coefficient) of the k-th component, where D1 = 0.5, D2 = 0.6, and D3 = 0.7; b k Let b be the k-th value, and the value of b is proportional to the square of the gradient, where b1 = 0, b2 = 3, and b3 = 6; Let S be the spectral matrix of the components, where the k-th row is the spectral signal of the k-th component, such as... Figure 3 As shown in (g).

[0056] S2, select the signal point of each spectral peak in the chemical shift dimension of the decay spectrum, and construct a feature vector corresponding to each signal point with a length equal to the number of gradients.

[0057] In a specific embodiment, step S2 specifically includes:

[0058] The chemical shift spectrum of the first gradient encoding is normalized, and data points with spectral intensities greater than a threshold are used as signal points.

[0059] Obtain the index value of the signal point in the attenuation spectrum, extract the corresponding column vector in the data matrix Y according to the index value, and normalize the column vector by the maximum value and take the natural logarithm to form a feature vector with a length equal to the number of gradients for each signal point.

[0060] Specifically, the signal points of each spectral peak along the chemical shift dimension, i.e., the signal points contained in each spectral peak, refer to... Figure 3 (b) and 3(c) normalize the chemical shift spectrum of the first gradient encoding, and use data points with spectral intensities greater than a threshold s as signal points. In this embodiment, s = 0.01 is selected. The thirty spectral peaks are labeled p1, p2, p3, ..., p1 from left to right. 30 A square pulse waveform marks signal points with spectral intensities greater than a threshold s (i.e., signal points contained within spectral peaks). Each signal point is modeled as a vector of length equal to the number of gradients, with vector elements representing the signal values ​​encoded by each gradient. Specifically, the index value of the signal point in the spectrum is obtained, and the corresponding column vector is extracted from the data matrix Y according to the index value. The column vector is then normalized to its maximum value and its natural logarithm is taken to form the feature vector c corresponding to the signal point. This embodiment uses three gradient codes, therefore each signal point has three feature values, is modeled as a vector of length 3, and each signal vector is normalized and its logarithm is taken. For clarity, in Figure 3 In (c), spectral peak p1 is magnified and displayed. v1, v2, v3, v4, and v5 are the five signal points contained within spectral peak p1, with index values ​​of 49–53 in the spectrum. Therefore, columns 49–53 of Y are taken, and the natural logarithm of each column is normalized to its maximum value to construct the eigenvectors c1, c2, c3, c4, and c5 corresponding to the five signal points, i.e.:

[0061] c1=lnY(:,49) / max(Y(:,49));

[0062] c2=lnY(:,50) / max(Y(:,50));

[0063] c3=lnY(:,51) / max(Y(:,51));

[0064] c4=lnY(:,52) / max(Y(:,52));

[0065] c5=lnY(:,53) / max(Y(:,53));

[0066] Here, ln(.) represents the natural logarithm function, Y(:,i) represents the i-th column vector of matrix Y, and max(.) is the maximum value function. Similarly, signal points are extracted sequentially for the remaining 29 spectral peaks, and corresponding eigenvectors are constructed. The length of this eigenvector is the number of gradients.

[0067] S3. Based on the feature vectors of all signal points of each spectral peak, a clustering algorithm is used to classify the signal points of each spectral peak to obtain the classification result of each spectral peak.

[0068] In specific embodiments, the clustering algorithms include CFSFDP clustering algorithm, K-Means clustering algorithm, K-Medoids clustering algorithm, agglomerative hierarchical clustering algorithm, and split hierarchical clustering algorithm.

[0069] In a specific embodiment, step S3 uses the CFSFDP clustering algorithm to classify all feature vectors according to their similarity to obtain the category of each signal point, where each category corresponds to a component.

[0070] Specifically, the cluster-based magnetic resonance diffusion-ordering spectroscopy processing method proposed in this application assumes, based on the differences in molecular self-diffusion properties, that gradient decay signals derived from the self-diffusion of the same molecule have greater similarity than decay signals derived from different molecules. This model transforms the separation problem of molecular component spectra in diffusion-ordering spectroscopy into a clustering problem, with each cluster corresponding to one molecule. Data points within the same cluster exhibit high similarity, while data points in different clusters show weaker similarity, thus achieving qualitative separation of molecular component spectra. The key to this method lies in using clustering algorithms to analyze diffusion-ordering spectral signal points, rather than the clustering algorithm itself. Therefore, this step can employ different clustering algorithms for analysis, including but not limited to CFSFDP clustering, K-Means clustering, K-Medoids clustering, agglomerative hierarchical clustering, and splitting hierarchical clustering.

[0071] Specifically, a clustering algorithm is used to classify all signal points based on their corresponding feature vectors c. In this embodiment, a density-based CFSFDP clustering algorithm (Clustering by Fast Search and Find of Density Peaks) is used to classify all feature vectors according to their similarity. Data within the same class have high similarity, while data between different classes have weak similarity. First, the center point of each class is found; the class center point is the point with the highest local density. After determining the class center point, the class is determined. The remaining points are assigned to the class belonging to the point with a higher density and the closest point to that point. In the embodiments of this application, reference is made to... Figure 3 (d) After analysis by the CFSFDP clustering algorithm, three classes were obtained, each corresponding to one component.

[0072] S4. Based on the classification results of the signal points of each spectral peak and the spectral intensity corresponding to each signal point, statistical analysis is performed to determine the category of each spectral peak. Based on the category, the magnetic resonance diffusion sorting spectrum is separated into components to obtain the separated spectrum corresponding to different components.

[0073] In a specific embodiment, step S4 specifically includes:

[0074] The classification weight vector of each spectral peak is obtained based on the category and spectral intensity of each signal point;

[0075] Each spectral peak is assigned a class based on the classification weight vector, thus obtaining the class of each spectral peak;

[0076] The magnetic resonance diffusion sorting spectrum is separated into components based on the category of each spectral peak to obtain the separated spectrum.

[0077] Specifically, the elements in the classification weight vector are the sum of the spectral intensities of all signal points under each category, and the category corresponding to the largest value among the elements of the classification weight vector is the class of each spectral peak.

[0078] Specifically, this step refers to classifying each spectral peak based on the signal point classification results. Therefore, different analytical methods with varying indices and standards can be used in this process. Specifically, statistical analysis can be performed on the classification results of the feature vectors corresponding to the signal points contained in the spectral peak to ultimately determine the class of each peak. The class of the spectral peak is voted on using the chemical shift intensity value corresponding to the signal point as a weight, ultimately determining the class of each peak. Here, each class corresponds to a molecular component, and the chemical shift spectral peaks are classified according to their molecular components, thereby achieving qualitative separation of each molecular component spectrum in the magnetic resonance diffusion ordering spectrum. The specific operation method is as follows: taking peak p1 as an example, refer to... Figure 3 (e) The spectral intensities corresponding to the five signal points v1, v2, v3, v4, and v5 of spectral peak p1 are w1 = 0.04, w2 = 0.06, w3 = 0.25, w4 = 0.05, and w5 = 0.02, respectively. The classification results include class 1, class 2, and class 3. Among them, signal point v5 belongs to class 1, signal point v1 belongs to class 2, and signal points v2, v3, and v4 belong to class 3. Therefore, the classification weight vector of p1 can be calculated. The i-th element of s1 represents the weight of the i-th class. Peak p1 is then assigned to class three, which has the highest weight. Each class corresponds to a component. By sequentially assigning the remaining chemical shift peaks to their respective classes using the same method, component separation of the diffusion-ordered spectrum can be achieved. The result is as follows: Figure 3 As shown in (f).

[0079] Ideal component separation reference spectrum as follows Figure 3 As shown in (g), the component separation spectrum obtained by the cluster-based magnetic resonance diffusion sorting spectroscopy processing method proposed in the embodiments of this application is as follows: Figure 4As shown in (f). Compared with the ideal component separation reference spectrum, the component separation spectrum obtained by the cluster-based magnetic resonance diffusion sorting spectrum processing method proposed in this application can accurately separate components under three gradients. Furthermore, the component separation spectrum obtained by the cluster-based magnetic resonance diffusion sorting spectrum processing method proposed in this application only took 4.8 seconds in this embodiment. It is evident that the component separation spectrum obtained by the cluster-based magnetic resonance diffusion sorting spectrum processing method proposed in this application can perform fast and accurate component separation of magnetic resonance diffusion sorting spectra under a small number of gradients, and has considerable application prospects.

[0080] Further reference Figure 2 As an implementation of the methods shown in the above figures, this application provides an embodiment of a cluster-based magnetic resonance diffusion sorting spectrum processing device, which is similar to... Figure 5 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0081] This application provides a cluster-based magnetic resonance diffusion sorting spectrum processing device, including:

[0082] Encoding module 1 is configured to acquire the attenuation spectrum corresponding to the magnetic resonance diffusion sorting spectrum with several gradient encodings;

[0083] Feature vector extraction module 2 is configured to select the signal point of each spectral peak in the chemical shift dimension of the decay spectrum and construct a feature vector corresponding to each signal point with a length equal to the number of gradients.

[0084] Clustering module 3 is configured to classify the signal points of each spectral peak using a clustering algorithm based on the feature vectors of all signal points of each spectral peak, and obtain the classification result of the signal points of each spectral peak.

[0085] Separation module 4 is configured to perform statistical analysis based on the classification results of the signal points of each spectral peak and the spectral intensity corresponding to each signal point, determine the category of each spectral peak, and separate the magnetic resonance diffusion sorting spectrum according to the category to obtain the separation spectrum corresponding to different components.

[0086] The following is for reference. Figure 1 It illustrates an electronic device suitable for implementing embodiments of this application (e.g., Figure 5 The diagram shows the structure of a computer device 500 (a server or terminal device). Figure 5 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0087] like ​As shown, the computer device 500 includes a central processing unit (CPU) 501 and a graphics processing unit (GPU) 502, which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 503 or programs loaded from storage section 509 into random access memory (RAM) 504. The RAM 504 also stores various programs and data required for the operation of the device 500. The CPU 501, GPU 502, ROM 503, and RAM 504 are interconnected via a bus 505. An input / output (I / O) interface 506 is also connected to the bus 505.

[0088] The following components are connected to I / O interface 506: an input section 507 including a keyboard, mouse, etc.; an output section 508 including an LCD, speakers, etc.; a storage section 509 including a hard disk, etc.; and a communication section 510 including a network interface card, such as a LAN card or modem. The communication section 510 performs communication processing via a network such as the Internet. A drive 511 may also be connected to I / O interface 506 as needed. A removable medium 512, such as a hard disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 511 as needed so that computer programs read from it can be installed into storage section 509 as needed.

[0089] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 510, and / or installed from removable medium 512. When the computer program is executed by central processing unit (CPU) 501 and graphics processing unit (GPU) 502, the functions defined in the methods of this application are performed.

[0090] It should be noted that the computer-readable medium described in this application can be a computer-readable signal medium, a computer-readable medium, or any combination thereof. A computer-readable medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor device, or any combination thereof. More specific examples of a computer-readable medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution device, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than a computer-readable medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution device, apparatus, or apparatus. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0091] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0092] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using dedicated hardware-based means to perform the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0093] The modules described in the embodiments of this application can be implemented in software or hardware. These modules can also be located within a processor.

[0094] In another aspect, this application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire attenuation spectra corresponding to several gradient-encoded magnetic resonance diffusion-ordered spectra; select signal points of each spectral peak in the chemical shift dimension of the attenuation spectrum and construct a feature vector corresponding to each signal point with a length equal to the number of gradients; classify the signal points of each spectral peak using a clustering algorithm based on the feature vectors of all signal points of each spectral peak to obtain the classification result of the signal points of each spectral peak; perform statistical analysis based on the classification result of the signal points of each spectral peak and the spectral intensity corresponding to each signal point to determine the category of each spectral peak; and separate the magnetic resonance diffusion-ordered spectra according to the category to obtain separated spectra corresponding to different components.

[0095] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A cluster-based magnetic resonance diffusion sorting spectrum processing method, characterized in that, Includes the following steps: S1, obtain the attenuation spectrum corresponding to the magnetic resonance diffusion sorting spectrum with several gradient encodings; S2, Select the signal point of each peak of the attenuation spectrum in the chemical shift dimension, and construct a feature vector corresponding to each signal point with a length equal to the number of gradients; S3. Based on the feature vectors of all signal points of each spectral peak, a clustering algorithm is used to classify the signal points of each spectral peak to obtain the classification result of each spectral peak. S4. Based on the classification results of the signal points of each spectral peak and the spectral intensity corresponding to each signal point, statistical analysis is performed to determine the category of each spectral peak. The magnetic resonance diffusion sorting spectrum is then separated into components according to the categories to obtain the separated spectra corresponding to different components. Specifically, this includes: The classification weight vector of each spectral peak is obtained based on the category and spectral intensity of each signal point; Each spectral peak is assigned a class based on the classification weight vector to obtain the class of each spectral peak. The elements in the classification weight vector are the sum of the spectral intensities of all signal points under each class, and the class corresponding to the largest value among the elements of the classification weight vector is the class of each spectral peak. The magnetic resonance diffusion sorting spectrum is component-separated according to the category of each spectral peak to obtain the separated spectrum.

2. The cluster-based magnetic resonance diffusion sorting spectrum processing method according to claim 1, characterized in that, In step S1, the attenuation spectrum is encoded using j gradients, the attenuation signal consists of i components, the chemical shift length is p, and the data matrix Y of the attenuation spectrum is represented as follows: ； in, , is the exponential decay matrix of the components. The Listed as the number The decay signal of each component; For the first The attenuation coefficient of each component; For the first indivual value, The value is proportional to the square of the gradient; , where is the spectral matrix of the components. The Behavior No. The spectral signals of each component.

3. The cluster-based magnetic resonance diffusion sorting spectrum processing method according to claim 2, characterized in that, Step S2 specifically includes: The chemical shift spectrum of the first gradient encoding is normalized, and data points with spectral intensities greater than a threshold are used as the signal points; Obtain the index value corresponding to the signal point in the attenuation spectrum, extract the corresponding column vector in the data matrix Y according to the index value, and normalize the column vector by the maximum value and take the natural logarithm to form the feature vector corresponding to each signal point.

4. The cluster-based magnetic resonance diffusion sorting spectrum processing method according to claim 1, characterized in that, The clustering algorithms include CFSFDP clustering algorithm, K-Means clustering algorithm, K-Medoids clustering algorithm, agglomerative hierarchical clustering algorithm, and split hierarchical clustering algorithm.

5. The cluster-based magnetic resonance diffusion sorting spectrum processing method according to claim 1, characterized in that, In step S3, the CFSFDP clustering algorithm is used to classify all feature vectors according to their similarity to obtain the category of each signal point, where each category corresponds to a component.

6. A cluster-based magnetic resonance diffusion sorting spectrum processing device, characterized in that, include: The encoding module is configured to acquire the attenuation spectrum corresponding to the magnetic resonance diffusion sorting spectrum with several gradient encodings. The feature vector extraction module is configured to select the signal point of each peak of the attenuation spectrum in the chemical shift dimension, and construct a feature vector corresponding to each signal point with a length equal to the number of gradients. The clustering module is configured to classify the signal points of each spectral peak using a clustering algorithm based on the feature vectors of all signal points of each spectral peak, and obtain the classification result of the signal points of each spectral peak. The separation module is configured to perform statistical analysis based on the classification result of the signal point of each spectral peak and the spectral intensity corresponding to each signal point to determine the category of each spectral peak, and to perform component separation of the magnetic resonance diffusion sorting spectrum according to the category to obtain the separation spectrum corresponding to different components, specifically including: The classification weight vector of each spectral peak is obtained based on the category and spectral intensity of each signal point; Each spectral peak is assigned a class based on the classification weight vector to obtain the class of each spectral peak. The elements in the classification weight vector are the sum of the spectral intensities of all signal points under each class, and the class corresponding to the largest value among the elements of the classification weight vector is the class of each spectral peak. The magnetic resonance diffusion sorting spectrum is component-separated according to the category of each spectral peak to obtain the separated spectrum.

7. An electronic device, comprising: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-5.

Images