CN113781365A - MPI system matrix restoration method based on depth image prior - Google Patents

MPI system matrix restoration method based on depth image prior Download PDF

Info

Publication number
CN113781365A
CN113781365A CN202111249534.4A CN202111249534A CN113781365A CN 113781365 A CN113781365 A CN 113781365A CN 202111249534 A CN202111249534 A CN 202111249534A CN 113781365 A CN113781365 A CN 113781365A
Authority
CN
China
Prior art keywords
system matrix
mpi
undersampled
rgb
matrix
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202111249534.4A
Other languages
Chinese (zh)
Inventor
田捷
尹琳
张鹏
惠辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Automation of Chinese Academy of Science
Original Assignee
Institute of Automation of Chinese Academy of Science
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute of Automation of Chinese Academy of Science filed Critical Institute of Automation of Chinese Academy of Science
Priority to CN202111249534.4A priority Critical patent/CN113781365A/en
Publication of CN113781365A publication Critical patent/CN113781365A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Epidemiology (AREA)
  • Medical Informatics (AREA)
  • Primary Health Care (AREA)
  • Public Health (AREA)
  • Image Processing (AREA)

Abstract

The invention belongs to the field of biomedical molecular images, and particularly relates to a method, a system and equipment for restoring an MPI system matrix based on depth image prior, aiming at solving the problems that the existing system matrix acquisition method is complicated in operation and long in acquisition time, and needs to be rescanned for new particles, so that the use efficiency of MPI is seriously influenced. The method comprises the following steps: constructing an undersampled system matrix; carrying out RGB encoding on the undersampled system matrix according to rows to obtain an RGB image; inputting random noise as network input into a 3D-Unet neural network, taking each RGB image as a truth label, iteratively calculating loss and returning each time, and outputting a fully sampled RGB image after iteration is finished; and decoding the fully sampled RGB image into a complex form to obtain a system matrix based on the complex number. The invention simplifies the acquisition operation of the system matrix, shortens the acquisition time and improves the use efficiency of MPI.

Description

MPI system matrix restoration method based on depth image prior
Technical Field
The invention belongs to the field of biomedical molecular images, and particularly relates to a method, a system and equipment for restoring an MPI system matrix based on depth image prior.
Background
Magnetic Particle Imaging (MPI) is an emerging tomographic technique that images Magnetic nanoparticles quantitatively in vivo with high spatial and temporal resolution. In addition, MPI only images magnetic nanoparticles specifically without interference of background signals, and therefore, the intensity of the acquired signal is proportional to the concentration of the magnetic nanoparticles, which makes MPI a very promising imaging modality in biomedical applications, such as visualization of cardiovascular diseases.
In order to reconstruct the spatial concentration distribution of the magnetic nanoparticles, a linear relationship between the measured voltage signal and the actual particle concentration distribution needs to be constructed with the aid of a system matrix. The system matrix describes a mapping between the concentration of particles and the induced voltage in the receiving coil. However, to date, the most accurate method of determining the matrix of an MPI system is to use a special calibration procedure. The basic idea is to fill a pixel or a grid on the grid with a small sample filled with magnetic nanoparticles based on the pre-defined three-dimensional grid data. This small sample was placed at all grid positions in the field of view (FOV) using a robot. At each location, MPI signal acquisition is performed, with the acquired signals as one column of the system matrix. The biggest disadvantage of the calibration method is the cumbersome operation and the long acquisition time, e.g. 34 x 28 x 20 grid images of medium size, which takes about 6 hours. In general, the higher the density of the grid, the higher the resolution of the image, and the more time it takes for the calibration process. Furthermore, due to operational variations in the magnetic nanoparticle synthesis process, the system matrix typically needs to be rescanned for each new batch of particles. Also any changes in the acquisition parameters, such as changes in the strength of the drive field or the selection field, require a new measurement system matrix. The problem seriously affects the use efficiency of the MPI and also greatly limits the application prospect of the MPI. Therefore, in order to further expand the application range of MPI imaging in biomedical research, a fast and accurate system matrix recovery method needs to be researched to accelerate the calibration process of the system matrix. Based on the method, the invention provides a depth image prior-based MPI system matrix restoration method.
Disclosure of Invention
In order to solve the above problems in the prior art, that is, to solve the problems that the existing system matrix acquisition method is complicated in operation and long in acquisition time, and new particles need to be rescanned, thereby seriously affecting the use efficiency of MPI and greatly limiting the application of MPI, the invention provides an MPI system matrix restoration method based on depth image prior, which restores a complete fully-sampled system matrix by using an under-sampled system matrix, further builds a linear relationship between a measurement voltage signal and actual particle concentration distribution based on the restored system matrix, reconstructs the spatial concentration distribution of magnetic nanoparticles, and performs magnetic particle imaging, the method comprises:
step S100, constructing an undersampled system matrix based on the length, width and height of the undersampled grid set in the MPI imaging process
Figure BDA0003322167950000021
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
s200, carrying out RGB encoding on the undersampled system matrix according to rows to obtain M RGB images;
step S300, inputting random noise as network input into a pre-constructed 3D-Unet neural network, taking each RGB image as a truth label, calculating loss and returning each time of iteration, and outputting a fully sampled RGB image after the iteration is finished;
step S400, decoding the fully sampled RGB image into a complex form to obtain a system matrix based on complex numbers, i.e., a restored system matrix.
In some preferred embodiments, the undersampled system matrix is constructed by:
according to the length, the width and the height of an undersampled grid set in MPI imaging, a robot automatically places a small sample filled with magnetic nanoparticles at each grid position of a visual field, MPI signal acquisition is carried out, voltage signals acquired at each position are used as a column of a system matrix, and after all grid positions are acquired, a corresponding undersampled system matrix is obtained.
In some preferred embodiments, the number of columns of the system matrix that is undersampled is obtained by:
N＝n1×n2×n3
wherein n is1，n2，n3Respectively showing the length, width and height of the set undersampling grid.
In some preferred embodiments, the under-sampled system matrix is RGB-encoded in rows to obtain M RGB images, and the method includes:
step S210: encoding each row vector of the undersampled system matrix into HSV format data; the row vector of the undersampled system matrix is in a complex form;
step S220: and converting the HSV data into an RGB image in a standard HSV-RGB conversion mode.
In some preferred embodiments, each row vector of the undersampled system matrix is encoded into HSV format data by:
THSV(Si)＝(H，S，V)＝(arg Si，1，|Si|)
wherein arg SiRepresenting the phase angle, | S of a complex numberiI denotes the magnitude of the complex number, i ═ 1.
In some preferred embodiments, the 3D-Unet neural network has a Loss function Loss during training as:
Figure BDA0003322167950000031
wherein, XoutRepresenting 3D-Unet neural networksThe fully sampled RGB image is output by the network and is subjected to set down sampling to obtain an under-sampled RGB image, XtrueThe RGB image obtained by RGB coding the matrix of the undersampled system according to rows is represented.
In a second aspect of the present invention, an MPI system matrix restoration system based on depth image prior is provided, which recovers a complete fully sampled system matrix by using an under-sampled system matrix, further constructs a linear relationship between a measurement voltage signal and an actual particle concentration distribution based on the recovered system matrix, reconstructs a spatial concentration distribution of magnetic nanoparticles, and performs magnetic particle imaging, and the system includes: the system comprises a system matrix construction module, an RGB coding module, an iterative computation module and a decoding module;
the system matrix construction module is configured to construct an undersampled system matrix based on the length, width and height of an undersampled grid set during MPI imaging
Figure BDA0003322167950000041
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
the RGB coding module is configured to carry out RGB coding on the under-sampled system matrix according to rows to obtain M RGB images;
the iteration calculation module is configured to input random noise as network input into a pre-constructed 3D-Unet neural network, take each RGB image as a truth label, calculate loss and return in each iteration, and output a fully sampled RGB image after the iteration is finished;
the decoding module is configured to decode the fully sampled RGB image into a complex form, and obtain a system matrix based on the complex number, that is, a restored system matrix.
In a third aspect of the present invention, an electronic device is provided, including: at least one processor; and a memory communicatively coupled to at least one of the processors; wherein the memory stores instructions executable by the processor for execution by the processor to implement the depth image prior-based MPI system matrix restoration method described above.
In a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions for being executed by the computer to implement the above-mentioned method for restoring a matrix of an MPI system based on a depth image prior.
The invention has the beneficial effects that:
the invention simplifies the acquisition operation of the system matrix, shortens the acquisition time and improves the use efficiency of MPI.
The invention does not need to train the network through a large amount of training data, but leads the network to continuously iteratively learn and update the network parameters based on the image data by embedding an image hard prior. The method solves the problems that the matrix of the MPI system is difficult to obtain and the data is seriously insufficient at present. The method improves the acquisition speed and the accuracy of the system matrix, makes up the defects of the traditional system matrix acquisition method based on calibration, improves the potential and the use efficiency of MPI imaging, and is convenient for the practical preclinical and clinical application of the subsequent MPI.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings.
FIG. 1 is a schematic flowchart of a depth image prior-based MPI system matrix restoration method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a framework of an MPI system matrix restoration system based on depth image priors according to an embodiment of the present invention;
FIG. 3 is a flowchart of a magnetic particle imaging method for MPI system matrix restoration based on depth image priors according to an embodiment of the present invention;
FIG. 4 is a block diagram of a magnetic particle imaging system for depth image prior-based MPI system matrix reconstruction according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a computer system suitable for implementing an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The present application will be described in further detail with reference to the following drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
The invention discloses a depth image prior-based MPI system matrix restoration method, which is characterized in that an under-sampled system matrix is used for restoring a complete full-sampled system matrix, a linear relation between a measurement voltage signal and actual particle concentration distribution is constructed based on the restored system matrix, the spatial concentration distribution of magnetic nanoparticles is reconstructed, and magnetic particle imaging is carried out, wherein the method comprises the following steps:
step S100, constructing an undersampled system matrix based on the length, width and height of the undersampled grid set in the MPI imaging process
Figure BDA0003322167950000061
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
s200, carrying out RGB encoding on the undersampled system matrix according to rows to obtain M RGB images;
step S300, inputting random noise as network input into a pre-constructed 3D-Unet neural network, taking each RGB image as a truth label, calculating loss and returning each time of iteration, and outputting a fully sampled RGB image after the iteration is finished;
step S400, decoding the fully sampled RGB image into a complex form to obtain a system matrix based on complex numbers, i.e., a restored system matrix.
In order to more clearly describe the depth image prior-based MPI system matrix restoration method of the present invention, the following describes in detail the steps in an embodiment of the method of the present invention with reference to the accompanying drawings.
The invention provides a depth image prior-based MPI system matrix restoration method, which can restore a fully sampled system matrix from an undersampled system matrix without training data or original truth data, so as to accelerate a calibration procedure of the system matrix and facilitate actual preclinical and clinical application of subsequent MPI. As shown in fig. 1, a depth image prior-based matrix restoration method for an MPI system specifically includes the following steps:
step S100, constructing an undersampled system matrix based on the length, width and height of the undersampled grid set in the MPI imaging process
Figure BDA0003322167950000071
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
in this embodiment, the calibration procedure is used to obtain an undersampled system matrix
Figure BDA0003322167950000072
For 3-dimensional MPI imaging, N ═ N1×n2×n3Wherein n is1，n2，n3Respectively representing the length, width and height of the set undersampling grid.
The calibration procedure is as follows:
according to the length, width and height of an undersampled grid set in MPI imaging, by means of a small sample filled with magnetic nanoparticles, a robot automatically places the small sample at each grid position of a field of view (FOV) according to a set program, and then MPI signal acquisition is carried out, wherein voltage signals acquired at each position are used as one column of a system matrix; after all grid positions are acquired, a corresponding undersampled system matrix is obtained.
S200, carrying out RGB encoding on the undersampled system matrix according to rows to obtain M RGB images;
in this embodiment, the under-sampled system matrix is RGB-encoded in rows to obtain M RGB images, and the method includes:
step S210: encoding each row vector (row vector is complex form) of the undersampled system matrix into HSV format data THSV(Si) The method comprises the following steps:
THSV(Si)＝(H，S，V)＝(arg Si，1，|Si|) (1)
wherein arg SiRepresenting the phase angle, | S of a complex numberiI denotes the magnitude of the complex number, i ═ 1.
Step S220: to hold HSV data THSV(Si) Converting into RGB image T in a standard HSV-RGB conversion modeRGB(Si)。
Step S300, inputting random noise as network input into a pre-constructed 3D-Unet neural network, taking each RGB image as a truth label, calculating loss and returning each time of iteration, and outputting a fully sampled RGB image after the iteration is finished;
in this embodiment, a 3D-Unet neural network is built, then random noise is input as network input to the built neural network, the RGB image obtained in step S200 is used as a true value for calculating loss and returning each iteration, and a fully sampled RGB image is output after the iteration is finished (in the present invention, the number of iterations is 2000). The loss function of the network training is a mean square error function, and the formula is as follows:
Figure BDA0003322167950000081
the output result of the 3D-Unet neural network is a full-sampling RGB image, and the truth label X of the networktrueAnd the RGB image obtained after the measured under-sampling system matrix is subjected to RGB coding according to rows is represented. Therefore, when calculating the loss, it is necessary to down-sample the fully sampled RGB image output from the network to obtain the restored under-sampled RGB image, i.e., Xout. Namely XoutIs represented by 3D-Unet neural network outputs fully sampled RGB image and performs set down sampling to obtain under-sampled RGB image, XtrueThe RGB image obtained by RGB coding the matrix of the undersampled system according to rows is represented.
Step S400, decoding the fully sampled RGB image into a complex form to obtain a system matrix based on complex numbers, i.e., a restored system matrix.
In this embodiment, the output result of the 3D-Unet neural network is RGB decoded to finally obtain a system matrix in a fully sampled complex form.
The method of the invention enables the network to carry out iterative learning based on the prior information of the image, and does not need any training data and training process. Compared with the traditional system matrix acquisition method based on calibration, the system matrix can be acquired quickly and accurately. And then constructing a linear relation between the measured voltage signal and the actual particle concentration distribution based on the recovered system matrix, reconstructing the spatial concentration distribution of the magnetic nanoparticles, performing magnetic particle imaging, and finally obtaining an MPI image.
A second embodiment of the present invention provides a depth image prior-based magnetic particle imaging method for matrix reconstruction of an MPI system, including:
step A100, constructing an undersampled system matrix based on the length, width and height of the undersampled grid set during MPI imaging
Figure BDA0003322167950000091
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
step A200, carrying out RGB encoding on the undersampled system matrix according to rows to obtain M RGB images;
step A300, inputting random noise as network input into a pre-constructed 3D-Unet neural network, taking each RGB image as a truth label, calculating loss and returning each time of iteration, and outputting a fully sampled RGB image after the iteration is finished;
step A400, decoding a fully sampled RGB image into a complex form to obtain a system matrix based on complex numbers, namely a restored system matrix;
and step A500, constructing a linear relation between the measured voltage signal and the actual particle concentration distribution based on the restored system matrix, reconstructing the spatial concentration distribution of the magnetic nanoparticles, and performing magnetic particle imaging.
Magnetic particle imaging based on the recovered system matrix is prior art and is not further described herein.
A third embodiment of the present invention provides a depth image prior-based MPI system matrix restoration system, which restores a complete fully sampled system matrix by using an undersampled system matrix, further constructs a linear relationship between a measurement voltage signal and an actual particle concentration distribution based on the restored system matrix, reconstructs a spatial concentration distribution of magnetic nanoparticles, and performs magnetic particle imaging, as shown in fig. 2, the system includes: the system comprises a system matrix construction module 100, an RGB coding module 200, an iterative computation module 300 and a decoding module 400;
the system matrix construction module 100 is configured to construct an under-sampled system matrix based on the length, width and height of an under-sampled grid set during MPI imaging
Figure BDA0003322167950000101
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
the RGB encoding module 200 is configured to perform RGB encoding on the under-sampled system matrix according to rows to obtain M RGB images;
the iterative computation module 300 is configured to input random noise as network input into a pre-constructed 3D-Unet neural network, use each RGB image as a true value label, compute loss and return each iteration, and output a fully sampled RGB image after the iteration is finished;
the decoding module 400 is configured to decode the fully sampled RGB image into a complex form, so as to obtain a system matrix based on the complex number, i.e. a recovered system matrix.
A fourth embodiment of the present invention provides a depth image prior-based magnetic particle imaging system for matrix reconstruction of an MPI system, the system comprising: the system comprises a system matrix construction module 10, an RGB coding module 20, an iterative computation module 30, a decoding module 40 and a magnetic particle imaging module 50;
the system matrix construction module 10 is configured to construct an under-sampled system matrix based on the length, width and height of an under-sampled grid set during MPI imaging
Figure BDA0003322167950000102
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
the RGB encoding module 20 is configured to perform RGB encoding on the undersampled system matrix according to rows to obtain M RGB images;
the iterative computation module 30 is configured to input random noise as network input to a pre-constructed 3D-Unet neural network, use each RGB image as a true value label, compute loss and return each iteration, and output a fully sampled RGB image after the iteration is finished;
the decoding module 40 is configured to decode the fully sampled RGB image into a complex form, so as to obtain a system matrix based on a complex number, that is, a restored system matrix;
the magnetic particle imaging module 50 is configured to construct a linear relationship between the measurement voltage signal and the actual particle concentration distribution based on the restored system matrix, reconstruct the spatial concentration distribution of the magnetic nanoparticles, and perform magnetic particle imaging.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process and related description of the system described above may refer to the corresponding process in the foregoing method embodiment, and details are not described herein again.
It should be noted that, the system for restoring an MPI system matrix based on depth image prior and/or the system for magnetic particle imaging based on restoration of an MPI system matrix based on depth image prior provided in the foregoing embodiments are only exemplified by the division of the above functional modules, and in practical applications, the above functions may be allocated to different functional modules according to needs, that is, the modules or steps in the embodiments of the present invention are decomposed or combined again, for example, the modules in the embodiments may be combined into one module, or may be further split into multiple sub-modules, so as to complete all or part of the above described functions. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps, and are not to be construed as unduly limiting the present invention.
An electronic device according to a fifth embodiment of the present invention includes at least one processor; and a memory communicatively coupled to at least one of the processors; wherein the memory stores instructions executable by the processor for execution by the processor to implement the method for depth image prior based MPI system matrix restoration and/or the method for magnetic particle imaging for depth image prior based MPI system matrix restoration of claims above.
A computer-readable storage medium of a sixth embodiment of the present invention stores computer instructions for execution by the computer to implement the method for restoration of a matrix of an MPI system based on depth image priors and/or the method for magnetic particle imaging for restoration of a matrix of an MPI system based on depth image priors of the claims above.
It can be clearly understood by those skilled in the art that, for convenience and brevity not described, the detailed working processes and related descriptions of the depth image prior based MPI system matrix restoration apparatus, the electronic device, and the computer-readable storage medium described above may refer to the corresponding processes in the foregoing method examples, and are not described herein again.
Referring now to FIG. 5, there is illustrated a block diagram of a computer system suitable for use as a server in implementing embodiments of the method, system, and apparatus of the present application. The server shown in fig. 5 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present application.
As shown in fig. 5, the computer system includes a Central Processing Unit (CPU) 501 that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 502 or a program loaded from a storage section 508 into a Random Access Memory (RAM) 503. In the RAM503, various programs and data necessary for system operation are also stored. The CPU501, ROM502, and RAM503 are connected to each other via a bus 504. An Input/Output (I/O) interface 505 is also connected to bus 504.
The following components are connected to the I/O interface 505: an input portion 506 including a keyboard, a mouse, and the like; an output section 507 including a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, a speaker, and the like; a storage portion 508 including a hard disk and the like; and a communication section 509 including a Network interface card such as a LAN (Local Area Network) card, a modem, or the like. The communication section 509 performs communication processing via a network such as the internet. The driver 510 is also connected to the I/O interface 505 as necessary. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 510 as necessary, so that a computer program read out therefrom is mounted into the storage section 508 as necessary.
In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 509, and/or installed from the removable medium 511. More specific examples of a computer-readable storage medium may include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), a compact disc read-only memory (CD-ROM), Optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. In the present application, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In this application, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for aspects of the present application may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The terms "first," "second," and the like are used for distinguishing between similar elements and not necessarily for describing or implying a particular order or sequence.
The terms "comprises," "comprising," or any other similar term are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
So far, the technical solutions of the present invention have been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of the present invention is obviously not limited to these specific embodiments. Equivalent changes or substitutions of related technical features can be made by those skilled in the art without departing from the principle of the invention, and the technical scheme after the changes or substitutions can fall into the protection scope of the invention.

Claims (9)

1. A MPI system matrix restoration method based on depth image prior utilizes an undersampled system matrix to restore a complete full-sampling system matrix, further constructs a linear relation between a measurement voltage signal and actual particle concentration distribution based on the restored system matrix, reconstructs the spatial concentration distribution of magnetic nanoparticles, and performs magnetic particle imaging, and is characterized in that the method comprises the following steps:
step S100, constructing an undersampled system matrix based on the length, width and height of the undersampled grid set in the MPI imaging process
Figure FDA0003322167940000011
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
s200, carrying out RGB encoding on the undersampled system matrix according to rows to obtain M RGB images;
step S300, inputting random noise as network input into a pre-constructed 3D-Unet neural network, taking each RGB image as a truth label, calculating loss and returning each time of iteration, and outputting a fully sampled RGB image after the iteration is finished;
step S400, decoding the fully sampled RGB image into a complex form to obtain a system matrix based on complex numbers, i.e., a restored system matrix.
2. The method for restoring the MPI system matrix based on the depth image priors according to claim 1, wherein the undersampled system matrix is constructed by:
according to the length, the width and the height of an undersampled grid set in MPI imaging, a robot automatically places a small sample filled with magnetic nanoparticles at each grid position of a visual field, MPI signal acquisition is carried out, voltage signals acquired at each position are used as a column of a system matrix, and after all grid positions are acquired, a corresponding undersampled system matrix is obtained.
3. The method for restoring the matrix of the MPI system based on the depth image priors as claimed in claim 1, wherein the number of columns of the system matrix under-sampled is obtained by:
N＝n1×n2×n3
wherein n is1，n2，n3Respectively showing the length, width and height of the set undersampling grid.
4. The method for restoration of the MPI system matrix based on depth image priors according to claim 1, wherein the under-sampled system matrix is RGB-encoded by rows to obtain M RGB images, and the method comprises:
step S210: encoding each row vector of the undersampled system matrix into HSV format data; the row vector of the undersampled system matrix is in a complex form;
step S220: and converting the HSV data into an RGB image in a standard HSV-RGB conversion mode.
5. The method for restoration of the MPI system matrix based on depth image priors of claim 4, wherein each row vector of the undersampled system matrix is encoded into HSV-format data by:
THSV(Si)＝(H，S，V)＝(arg Si，1，|Si|)
wherein arg SiRepresenting the phase angle, | S of a complex numberiI denotes the magnitude of the complex number, i 1.
6. The method for restoring the matrix of the MPI system based on the depth image priors as claimed in claim 1, wherein the Loss function Loss of the 3D-Unet neural network during training is:
Figure FDA0003322167940000021
wherein, XoutThe RGB image which represents the full sampling output by the 3D-Unet neural network is subjected to set down-sampling to obtain under-samplingRGB image of (1), XtrueThe RGB image obtained by RGB coding the matrix of the undersampled system according to rows is represented.
7. The utility model provides a MPI system matrix system of restoreing based on depth map priori for obtain the system matrix after restoreing, and then based on the system matrix after restoreing, construct the linear relation between measurement voltage signal and the actual particle concentration distribution, rebuild the space concentration distribution of magnetic nanoparticle, carry out magnetic particle imaging, its characterized in that, this system includes: the system comprises a system matrix construction module, an RGB coding module, an iterative computation module and a decoding module;
the system matrix construction module is configured to construct an undersampled system matrix based on the length, width and height of an undersampled grid set during MPI imaging
Figure FDA0003322167940000031
Wherein M, N represents the number of rows and columns of the undersampled system matrix;
the RGB coding module is configured to carry out RGB coding on the under-sampled system matrix according to rows to obtain M RGB images;
the iteration calculation module is configured to input random noise as network input into a pre-constructed 3D-Unet neural network, take each RGB image as a truth label, calculate loss and return in each iteration, and output a fully sampled RGB image after the iteration is finished;
the decoding module is configured to decode the fully sampled RGB image into a complex form, and obtain a system matrix based on the complex number, that is, a restored system matrix.
8. An electronic device, comprising:
at least one processor; and a memory communicatively coupled to at least one of the processors;
wherein the memory stores instructions executable by the processor for execution by the processor to implement the depth image prior-based MPI system matrix restoration method of any one of claims 1-6.
9. A computer-readable storage medium storing computer instructions for execution by the computer to implement the depth image prior-based MPI system matrix restoration method of any one of claims 1 to 6.
CN202111249534.4A 2021-10-26 2021-10-26 MPI system matrix restoration method based on depth image prior Pending CN113781365A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111249534.4A CN113781365A (en) 2021-10-26 2021-10-26 MPI system matrix restoration method based on depth image prior

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111249534.4A CN113781365A (en) 2021-10-26 2021-10-26 MPI system matrix restoration method based on depth image prior

Publications (1)

Publication Number Publication Date
CN113781365A true CN113781365A (en) 2021-12-10

Family

ID=78956636

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111249534.4A Pending CN113781365A (en) 2021-10-26 2021-10-26 MPI system matrix restoration method based on depth image prior

Country Status (1)

Country Link
CN (1) CN113781365A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114581553A (en) * 2022-04-28 2022-06-03 北京航空航天大学 Fluorescent molecular tomography reconstruction method based on magnetic particle imaging prior guidance
CN114998471A (en) * 2022-06-22 2022-09-02 中国科学院自动化研究所 Magnetic particle imaging reconstruction method based on RecNet model

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114581553A (en) * 2022-04-28 2022-06-03 北京航空航天大学 Fluorescent molecular tomography reconstruction method based on magnetic particle imaging prior guidance
CN114581553B (en) * 2022-04-28 2022-07-22 北京航空航天大学 Fluorescent molecular tomography reconstruction method based on magnetic particle imaging prior guidance
CN114998471A (en) * 2022-06-22 2022-09-02 中国科学院自动化研究所 Magnetic particle imaging reconstruction method based on RecNet model
CN114998471B (en) * 2022-06-22 2022-12-16 中国科学院自动化研究所 Magnetic particle imaging reconstruction method based on RecNet model
US11835602B1 (en) 2022-06-22 2023-12-05 Institute Of Automation, Chinese Academy Of Sciences Magnetic particle imaging (MPI) reconstruction method based on RecNet model

Similar Documents

Publication Publication Date Title
CN113781365A (en) MPI system matrix restoration method based on depth image prior
CN109239631B (en) Magnetic resonance imaging method and device
WO2020114329A1 (en) Fast magnetic resonance parametric imaging and device
US11763134B2 (en) Systems and methods for image reconstruction
CN116503507B (en) Magnetic particle image reconstruction method based on pre-training model
CN114998471B (en) Magnetic particle imaging reconstruction method based on RecNet model
CN109146787A (en) A kind of real-time reconstruction method of the double camera spectrum imaging system based on interpolation
CN114723729A (en) MPI image reconstruction method based on Transformer
CN115797493B (en) Magnetic field free line magnetic particle imaging method based on one-dimensional system matrix sparse sampling
CN112017161A (en) Pulmonary nodule detection method and device based on central point regression
CN115952743A (en) Multi-source precipitation data collaborative downscaling method and system coupled with random forest and HASM
WO2021129235A1 (en) Rapid three-dimensional magnetic resonance parameter imaging method and device
TWI478103B (en) Diffusion spectrum image transforming method of using large deformation diffeomorphic metric mapping
WO2022193378A1 (en) Image reconstruction model generation method and apparatus, image reconstruction method and apparatus, device, and medium
CN116934597B (en) FDNN model-based magnetic particle imaging spatial resolution improvement method
CN115778353B (en) Magnetic field free line magnetic particle imaging method based on rotation harmonic diagram
Linnér et al. LatticeLibrary and BccFccRaycaster: software for processing and viewing 3D data on optimal sampling lattices
JP2009524070A (en) Improved interpolation method between multiple observed tensors
CN115356672B (en) Multi-dimensional magnetic resonance imaging method, system and storage medium
CN116740217B (en) Arterial spin marking method, device and storage medium based on artificial intelligence technology
CN117689761B (en) Plug-and-play magnetic particle imaging reconstruction method and system based on diffusion model
CN115272506A (en) Image reconstruction method of large-aperture magnetic particle imaging system based on multiple prior characteristics
CN114820329B (en) Curved surface measuring method and device based on Gaussian process large-kernel attention device guidance
US20220343634A1 (en) Pseudo data generation apparatus, pseudo data generation method, and non-transitory storage medium
CN114881208A (en) MMPI mixed signal separation method based on self-attention mechanism

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination