Image reconstruction method, device, electronic device, storage medium and computer program

Abstract

The present application relates to an image reconstruction method, device, electronic device, storage medium and computer program. The method includes: acquiring multiple K-space data sets of a region of interest, each K-space data set corresponding to a coil unit in a magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit. An initial image corresponding to the coil unit is obtained based on the K-space data set. The initial image is weighted filtered to obtain a filtered image corresponding to the initial image. A sensitivity estimate of each coil unit is obtained. Based on the sensitivity estimate of each coil unit and a sensitivity parallel imaging algorithm, the filtered image corresponding to each coil unit is reconstructed to obtain a target image. This method can solve the problem of poor noise suppression in the existing image reconstruction process, resulting in low quality of the reconstructed image.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to an image reconstruction method, device, electronic device, storage medium and computer program. Background Art

[0002] Magnetic resonance imaging (MRI) technology is currently widely used in clinical diagnosis and analysis. This technology uses mathematical methods to reconstruct the signals generated by nuclear resonance to generate images of the internal structures of the human body.

[0003] During MRI image acquisition, curling artifacts may appear in the image to increase imaging speed. MRI reconstruction can reduce or eliminate curling artifacts in MRI images, yielding a "true" original image. However, due to the significant amount of noise associated with curling artifacts, related techniques are ineffective in suppressing noise during image reconstruction, resulting in lower reconstructed image quality. Summary of the Invention

[0004] Based on this, it is necessary to provide an image reconstruction method, device, electronic device, storage medium and computer program to address the above technical problems, so as to solve the problem that the noise suppression effect in the existing image reconstruction process is poor, resulting in low quality of the reconstructed image.

[0005] In a first aspect, the present application provides an image reconstruction method, the method comprising:

[0006] A plurality of K-space data sets of a region of interest are acquired, each K-space data set corresponding to a coil unit in a magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0007] An initial image corresponding to the coil unit is obtained according to the K-space data set.

[0008] Perform weighted filtering on the initial image to obtain a filtered image corresponding to the initial image.

[0009] Get the sensitivity estimate of each coil unit.

[0010] According to the sensitivity estimation of each coil unit and the sensitivity parallel imaging algorithm, image reconstruction is performed on the filtered image corresponding to each coil unit to obtain a target image.

[0011] In one embodiment, performing weighted filtering on the initial image to obtain a filtered image corresponding to the initial image includes:

[0012] A filtering parameter of the coil unit is determined.

[0013] Based on the filtering parameters, weighted filtering is performed on the initial image corresponding to the coil unit to obtain a filtered image corresponding to the initial image.

[0014] In one embodiment, determining the filtering parameters of the coil unit includes:

[0015] An initial noise variance level of the initial image is determined according to the initial image of the coil unit.

[0016] The initial noise variance level is normalized to obtain the target noise variance.

[0017] A filtering parameter of the coil unit is determined according to the target noise variance level.

[0018] In one embodiment, determining the filtering parameters of the coil unit according to the target noise variance level includes:

[0019] The target noise variance level is matched with a preset condition to obtain a window size and a number of iterations in the filter parameters of the coil unit.

[0020] In one embodiment, determining the filtering parameters of the coil unit according to the target noise variance level includes:

[0021] The weights of the filter parameters of the coil unit are obtained by calculation according to the target noise variance level.

[0022] In one embodiment, the weight is any one of a Gaussian weight and a reciprocal weight.

[0023] In a second aspect, the present application further provides an image reconstruction device, comprising:

[0024] The data acquisition module is used to acquire multiple K-space data sets of the region of interest, each K-space data set corresponds to a coil unit in the magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0025] The initial image acquisition module is used to obtain an initial image corresponding to the coil unit according to the K-space data set acquired by the data acquisition module.

[0026] The weighted filtering processing module is used to perform weighted filtering processing on the initial image obtained by the initial image acquisition module to obtain a filtered image corresponding to the initial image.

[0027] The image reconstruction module is used to obtain the sensitivity estimation of each coil unit.

[0028] The image reconstruction module is further configured to reconstruct the filtered image corresponding to each coil unit based on the sensitivity estimation of each coil unit obtained by the weighted filtering processing module and the sensitivity parallel imaging algorithm to obtain a target image.

[0029] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0030] A plurality of K-space data sets of a region of interest are acquired, each K-space data set corresponding to a coil unit in a magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0031] An initial image corresponding to the coil unit is obtained according to the K-space data set.

[0032] Perform weighted filtering on the initial image to obtain a filtered image corresponding to the initial image.

[0033] Get the sensitivity estimate of each coil unit.

[0034] According to the sensitivity estimation of each coil unit and the sensitivity parallel imaging algorithm, image reconstruction is performed on the filtered image corresponding to each coil unit to obtain a target image.

[0035] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0036] A plurality of K-space data sets of a region of interest are acquired, each K-space data set corresponding to a coil unit in a magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0037] An initial image corresponding to the coil unit is obtained according to the K-space data set.

[0038] Perform weighted filtering on the initial image to obtain a filtered image corresponding to the initial image.

[0039] Get the sensitivity estimate of each coil unit.

[0040] According to the sensitivity estimation of each coil unit and the sensitivity parallel imaging algorithm, image reconstruction is performed on the filtered image corresponding to each coil unit to obtain a target image.

[0041] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps:

[0042] A plurality of K-space data sets of a region of interest are acquired, each K-space data set corresponding to a coil unit in a magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0043] An initial image corresponding to the coil unit is obtained according to the K-space data set.

[0044] Perform weighted filtering on the initial image to obtain a filtered image corresponding to the initial image.

[0045] Get the sensitivity estimate of each coil unit.

[0046] According to the sensitivity estimation of each coil unit and the sensitivity parallel imaging algorithm, image reconstruction is performed on the filtered image corresponding to each coil unit to obtain a target image.

[0047] The above-described image reconstruction method, apparatus, electronic device, storage medium, and computer program obtain an initial image of the coil unit based on a K-space dataset obtained by undersampling the coil units in the magnetic resonance coil. The initial image is then subjected to weighted filtering to obtain a filtered image corresponding to the initial image. The initial image is then reconstructed using the sensitivity estimates for each coil unit to obtain a target image. By performing weighted filtering on each initial image to remove noise from the initial image, the problem of poor noise suppression in existing image reconstruction processes can be addressed. Furthermore, the reconstructed image is obtained by combining the noise-filtered initial image with the sensitivity estimates, thereby improving the quality of the reconstructed image. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is an architectural diagram of an image reconstruction system in one embodiment;

[0049] Figure 2 FIG1 is a flow chart of an image reconstruction method according to an embodiment;

[0050] Figure 3 is a schematic diagram of the structure of a K-space data set in one embodiment;

[0051] Figure 4 Schematic diagram of the effect of an image reconstruction method in one embodiment;

[0052] Figure 5 FIG2 is a second flow chart of an image reconstruction method according to an embodiment;

[0053] Figure 6FIG3 is a flowchart of an image reconstruction method according to an embodiment;

[0054] Figure 7 is a schematic structural diagram of an image reconstruction device in one embodiment;

[0055] Figure 8 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0056] The following will be combined with the accompanying drawings to clearly and completely describe the technical solutions in some embodiments of the present application. Obviously, the embodiments described are only some embodiments of the present application, not all embodiments. Based on the embodiments provided in this application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0057] Unless the context requires otherwise, throughout the specification and claims, the term "comprise" and its other forms, such as the third person singular form "comprises" and the present participle form "comprising", are to be interpreted as open and inclusive, that is, "including, but not limited to". In the description of the specification, the terms "one embodiment", "some embodiments", "exemplary embodiments", "example", "specific example" or "some examples" are intended to indicate that the specific features, structures, materials or characteristics associated with the embodiment or example are included in at least one embodiment or example of the present application. The schematic representation of the above terms does not necessarily refer to the same embodiment or example. In addition, the specific features, structures, materials or characteristics may be included in any one or more embodiments or examples in any appropriate manner.

[0058] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of this application, unless otherwise specified, "plurality" means two or more.

[0059] When describing some embodiments, the expressions "coupled" and "connected" and their derivatives may be used. For example, when describing some embodiments, the term "connected" may be used to indicate that two or more components are in direct physical or electrical contact with each other. For another example, when describing some embodiments, the term "coupled" may be used to indicate that two or more components are in direct physical or electrical contact. However, the term "coupled" or "communicatively coupled" may also refer to two or more components that are not in direct contact with each other, but still cooperate or interact with each other. The embodiments disclosed herein are not necessarily limited to the contents of this document.

[0060] “At least one of A, B and C” has the same meaning as “at least one of A, B or C” and both include the following combinations of A, B and C: A only, B only, C only, the combination of A and B, the combination of A and C, the combination of B and C, and the combination of A, B and C.

[0061] As used herein, the term "if" is optionally interpreted to mean "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined that" or "if [stated condition or event] is detected" are optionally interpreted to mean "upon determining" or "in response to determining" or "upon detecting [stated condition or event]" or "in response to detecting [stated condition or event]," depending on the context.

[0062] The use of "adapted to" or "configured to" herein is intended to be open and inclusive language that does not exclude devices adapted or configured to perform additional tasks or steps.

[0063] Additionally, the use of “based on” or “according to” is meant to be open and inclusive, as a process, step, calculation, or other action “based on” or “according to” one or more stated conditions or values ​​may, in practice, be based on additional conditions or values ​​beyond those stated.

[0064] Multi-coil imaging technology is a new magnetic resonance imaging technology developed in recent years. It is based on a multi-channel phased array coil and has the characteristics of high signal-to-noise ratio and high image spatial resolution. Usually, a phased array coil is a coil array composed of two or more coil units, and each coil unit can receive signals from various areas at the same time. The magnetic resonance signal obtained by the multi-channel radiofrequency (RF) coil as a whole is the effective weighted sum of the signals obtained by each coil unit, but the noise contained in it only comes from the small area defined by each coil unit. Therefore, the image obtained using a multi-channel RF coil has a higher signal-to-noise ratio. Multi-coil imaging combines the advantages of small coil imaging with the large scanning field of view of large coil imaging.

[0065] To increase image acquisition speed, an acceleration factor is added to each coil unit when collecting image data. This results in missing K-space rows, resulting in undersampled K-space datasets. This causes each coil-reconstructed image to contain warping artifacts, which in turn contain significant noise. Existing techniques for processing K-space datasets include multiplication by a window function or low-pass filtering to suppress high-spatial frequency signals, such as Gibbs artifacts. However, these methods are less effective in suppressing noise in the image, resulting in lower reconstructed image quality.

[0066] To address the above technical issues, an embodiment of the present application provides an image reconstruction method. This method obtains an initial image of each coil unit based on a K-space dataset of coil units in a magnetic resonance coil obtained through undersampling. Based on each initial image, the corresponding noise variance level is determined. The noise variance level is used to determine weighted filtering parameters for the initial image, thereby obtaining a filtered image after filtering the initial image. Furthermore, based on the filtered images corresponding to each coil unit and the sensitivity estimates for each coil unit, a target image is obtained. This ensures noise suppression during the image reconstruction process and improves the quality of the reconstructed image.

[0067] For ease of use of this embodiment, see Figure 1 The image reconstruction system 10 shown in FIG. 1 includes an image reconstruction device 11 and an imaging device 12. The imaging device 12 can capture a region of interest and obtain a K-space data set of a coil unit.

[0068] In an exemplary solution, under normal circumstances, the image reconstruction device 11 can be a terminal device; the terminal device can have a general or special computing device environment or configuration. For example: a personal computer, a server computer, a handheld device or a portable device, a tablet device, a multi-processor device, a distributed computing environment including any of the above devices or devices, etc. The terminal device can have different names, such as user equipment (UE), access device, terminal unit, terminal station, mobile station, mobile station, remote station, remote terminal, mobile device, wireless communication device, terminal agent or terminal device, etc. In the embodiment of the present application, the device for realizing the function of the image reconstruction device 11 can be a terminal device, or it can be a device that can support the image reconstruction device 11 to realize the function, such as a chip system, etc. In the present application, the chip system can be composed of a chip, or it can include chips and other discrete devices.

[0069] Combined with the above Figure 1 , the image reconstruction method provided in the embodiment of the present application is described in detail, referring to Figure 2 , the method comprising:

[0070] S11. Acquire multiple K-space data sets of a region of interest, where each K-space data set corresponds to a coil unit in the magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0071] Here, under-sampled data specifically refers to adding an acceleration factor when the coil unit in the magnetic resonance coil collects image data.

[0072] Reference Figure 3 The present invention provides a structure of a K-space data set. K-space is an abstract space, which is a frequency space corresponding to a spatial coordinate system with spatial frequency as the unit. The grayscale value of each pixel in K-space determines the intensity of the magnetic resonance signal at that point, and its position value is determined by the phase encoding gradient G. y and frequency encoding gradient G x The data for each point in k-space is sequentially filled along a specific trajectory, representing the acquisition process of the magnetic resonance signal during imaging. Conventional magnetic resonance imaging mostly uses a linear filling method, filling one row per phase encoding cycle until the data for each point in k-space is completely filled.

[0073] In practical applications, to increase the speed of image acquisition, an acceleration factor is added to the coil unit when collecting image data. The acceleration factor is used to determine the number of phase encoding lines required for the K-space sampling trajectory. Generally, this can be set at the initialization stage. For example, if the acceleration factor is initially set to 1, it indicates that the K-space sampling trajectory is fully sampled; for example, if the acceleration factor is initially set to 2, it indicates that the K-space sampling trajectory is undersampled. K-space can include low-frequency regions and high-frequency regions. After the low-frequency region is acquired according to the adjusted acceleration factor, the high-frequency region must also acquire phase encoding lines according to the adjusted acceleration factor to achieve the set number of phase encoding lines.

[0074] In one example, magnetic resonance (MR) image acquisition involves acquiring and filling K-space with phase-encoding lines. The more phase-encoding lines required to acquire K-space, the longer the image acquisition time. Therefore, the density of the phase-encoding lines determines the image's field of view (FOV) in the phase-encoding direction. Using a rectangular FOV technique can reduce the number of phase-encoding lines required, sparsely filling the entire K-space. This proportionally shortens acquisition time while maintaining image spatial resolution, but reduces the image's FOV in the phase-encoding direction. For example, for a 256×256 pixel image, if only 50% (i.e., 128) of the phase-encoding lines are acquired and filled throughout K-space, the acquisition time is halved, and the image's FOV in the phase-encoding direction is reduced to 50%. Using multiple coils to acquire MR signals, each with a specific sensitivity, and with varying sensitivities across coil units, allows the acquired MR signals to contain additional spatial information.

[0075] S12. Obtain an initial image corresponding to the coil unit according to the K-space data set.

[0076] Specifically, an inverse Fourier transform is performed on the K-space data set to obtain an initial image of each coil.

[0077] Here, the magnetic resonance coil can be a coil array comprising multiple coil units. The magnetic resonance signals acquired by the coil units can be combined into multiple K-space datasets; wherein, the K-space dataset of each coil can be Fourier transformed to obtain an image corresponding to that K-space. For example, when a multi-channel RF coil includes a Q channel, Q K-spaces can be obtained by filling in phase encoding lines. These Q K-spaces can be Fourier transformed to obtain Q images, where each image corresponds to a signal acquisition channel.

[0078] It should be noted that the inverse Fourier transform referred to in this application may also be a fast inverse Fourier transform. Performing an inverse Fourier transform on a K-space dataset refers to transforming the K-space dataset domain into the image domain. For example, the signal of a multi-channel image and the K-space dataset have the following relationship:

[0079]

[0080] in, Indicates the coil unit in K space (k x ,k y ), and the K-space dataset is undersampled and may also be filtered; S (x,y) Represents the image signal of the coil unit at the image domain position (x, y).

[0081] It should be understood that the undersampled k-space dataset is converted from k-space to image space using inverse fast Fourier transform. Since inverse fast Fourier transform requires less computation, the use of inverse fast Fourier transform enables reducing the acquisition time of magnetic resonance images.

[0082] S13: Perform weighted filtering on the initial image to obtain a filtered image corresponding to the initial image.

[0083] In one possible implementation, weighted filtering of the initial image can be understood as a smoothing process (smoothing) of the initial image, also known as blurring. Its function is to reduce noise or distortion on the image. From the perspective of signal processing, image smoothing is to remove high-frequency information and retain low-frequency information. Therefore, low-pass filtering can be performed on the image. Low-pass filtering can remove noise in the image and blur the image (noise is the area with relatively large changes in the image, that is, high-frequency information). High-pass filtering can extract the edges of the image (edges are also areas where high-frequency information is concentrated). The smoothing process can be achieved by a filter, which can be a weighted filter;

[0084] Furthermore, the initial image is subjected to weighted filtering to obtain a filtered image corresponding to the initial image. Any one of filters such as a mean filter, a Gaussian weighted filter, a median filter, a bilateral filter, etc. may be used, and this is not limited in the embodiment of the present application.

[0085] S14. Obtain sensitivity estimation of each coil unit.

[0086] Specifically, the process of obtaining sensitivity estimates for each coil unit includes: acquiring a low-resolution volume transmit coil (VTC) image (i.e., a reference image) and a low-resolution surface coil image; and dividing the VTC image by the surface coil image (removing anatomical structures while retaining coil sensitivity information) to obtain an initial sensitivity estimate. Considering that the signal-to-noise ratio (SNR) of the initial sensitivity estimate obtained by directly dividing the VTC image by the surface coil image is very low, mainly due to the high image noise from the volume coil acquisition, a polynomial fitting method can be used to smooth the noise. Furthermore, the relative movement of the measured object can cause edge errors. Therefore, the initial sensitivity estimate can be Gaussian smoothed to remove noise and obtain the sensitivity estimate for each coil unit.

[0087] S15 . Reconstruct the filtered images corresponding to the coil units according to the sensitivity estimation of the coil units and the sensitivity parallel imaging algorithm to obtain a target image.

[0088] It should be noted that the magnetic resonance coil in this application may be an array coil having multiple coil units, i.e., multiple acquisition channels. Each channel of the array coil corresponds to a sensitivity estimate, and each channel of the array coil needs to obtain a corresponding sensitivity estimate.

[0089] In one embodiment, based on the sensitivity estimation of each coil unit and the sensitivity coding parallel imaging algorithm, the overlapping parts of the filtered images corresponding to each coil unit are expanded and combined to obtain the target image.

[0090] Here, sensitivity encoding (SENSE) parallel magnetic resonance imaging technology (i.e., sensitivity encoding parallel imaging algorithm) is a typical image domain reconstruction algorithm. Its reconstruction steps mainly include: (1) Undersampling the multiple channels corresponding to the multiple coil units with an acceleration factor R to obtain the K-space data set of each channel, performing inverse Fourier transform on the K-space data set obtained by each coil unit to obtain the initial image of each coil unit, and performing weighted filtering on the initial image to obtain a filtered image; wherein the filtered image can be understood as a convoluted image; (2) Analyzing the sensitivity distribution of the parallel coil units, unfolding the convoluted image on the sensitivity distribution map (i.e., unfolding and combining the overlapping parts of the filtered images corresponding to each coil unit), thereby obtaining a full-field-of-view (FOV) image, i.e., a complete reconstructed image (i.e., the target image).

[0091] Specifically, SENSE is a parallel imaging technology. The relationship between the sensitivity estimate S of the coil unit, the filtered image I of the coil unit, and the target image P is:

[0092] I = S × P (Formula 2)

[0093] The process of reconstructing the image is actually the inverse process of the above formula 2, which is:

[0094] P=S -1 ×I (Formula 3)

[0095] In this embodiment, based on the sensitivity estimation and sensitivity encoding SENSE parallel imaging algorithm of each coil unit, the overlapping parts of the filtered images corresponding to each coil unit are expanded and combined to obtain the target image, which can reduce the noise in the target image and remove the curling artifacts in the target image, thereby improving the clarity of the reconstructed image.

[0096] For better understanding, refer to Figure 4First, each coil unit (Coil 1, Coil 2, ..., Coil k) acquires its own K-space dataset, and an initial image is generated from this K-space dataset. Either an inverse Fourier transform or a fast Fourier transform can be performed on the K-space dataset to obtain the corresponding initial image. Next, a weighted filter is applied to the initial image to obtain a filtered image. This filtered image is then reconstructed using the SENSE algorithm based on the sensitivity estimates of each coil unit to obtain the target image.

[0097] The above-described image reconstruction method obtains an initial image of the coil units based on a K-space dataset acquired by undersampling the coil units in the magnetic resonance coil. The initial image is then subjected to weighted filtering to obtain a filtered image corresponding to the initial image. The initial image is then reconstructed using the sensitivity estimates for each coil unit to obtain the target image. By performing weighted filtering on each initial image to remove noise from the initial image, the problem of poor noise suppression in existing image reconstruction processes can be addressed. Furthermore, the reconstructed image is obtained by combining the noise-filtered initial image with the sensitivity estimates, thereby improving the quality of the reconstructed image.

[0098] In one embodiment, referring to Figure 5 In the process of weighted filtering the initial image corresponding to each coil unit, the initial image may be weighted filtered according to the filtering parameters of each coil unit. Therefore, S13 specifically includes:

[0099] S131. Determine filtering parameters of the coil unit.

[0100] Optionally, the filtering parameters include weight, filter window size, and number of iterations.

[0101] Here, the weight is either Gaussian weight or reciprocal weight.

[0102] Specifically, when the weight is a Gaussian weight, the weight u can be calculated as follows:

[0103]

[0104] Among them, F(q) is the signal strength corresponding to the center point q of the filter window; F(p) represents the signal strength corresponding to other points p in the filter window except the center point q. represents the smoothing parameter of the filter, The value of is obtained according to the target noise variance level, The larger the value, the smoother the filter. The filter's input parameters include the filter window size w and the number of iterations iter. The number of iterations indicates the number of times the filter is repeated. The larger the w and iter are, the smoother the filter is.

[0105] When the weight is a reciprocal weight, the reciprocal weight is further divided into Type I weight and Type II weight.

[0106] When the reciprocal weight is type I weight, the weight u can be calculated as follows:

[0107] u=(|F(p)-F(q)|+δ f ) -1 (Formula 4)

[0108] When the reciprocal weight is a type II weight, the weight u can be calculated as follows:

[0109]

[0110] It should be noted that the weighted filters, mean filters, Gaussian weighted filters, median filters, and bilateral filters mentioned above can have adjustable weight types. For example, for weighted filters, either the Gaussian weight or the reciprocal weight can be selected as the weight of the weighted filter; for Gaussian weighted filters, the Gaussian weight can be used as the weight of the Gaussian weighted filter; the same applies to other filters and will not be described here.

[0111] S132 . Perform weighted filtering on the initial image corresponding to the coil unit based on the filtering parameters to obtain a filtered image corresponding to the initial image.

[0112] For the embodiment of this step, reference may be made to the embodiment of S13, which will not be described in detail here.

[0113] In this embodiment, based on the filtering parameters of the coil unit, the initial image corresponding to the coil unit is subjected to weighted filtering processing to obtain a filtered image corresponding to the initial image. This can improve the noise suppression effect in the image to be reconstructed, reduce the impact of noise on the image to be reconstructed, and make the reconstructed image clearer.

[0114] In one embodiment, referring to Figure 6 , determine the filtering parameters of the coil unit, including:

[0115] S61 . Determine an initial noise variance level of the initial image according to the initial image of the coil unit.

[0116] Specifically, according to the initial image of the coil unit, the serial number k of the coil unit and the element IM in the i-th row and j-th column of the initial image of the k-th coil unit are determined. k (i, j), the initial noise variance level is determined according to the following formula 5

[0117]

[0118] S62, normalizing the initial noise variance level to obtain a target noise variance;

[0119] Specifically, the kth initial noise variance level calculated from the initial image of the kth coil unit is normalized to obtain the target noise variance level of the kth coil unit. It should be noted here that in order to distinguish the initial noise variance level from the target noise variance level, k is replaced by i, and k and i have the same meaning.

[0120] S63. Determine the filtering parameters of the coil unit according to the target noise variance level.

[0121] Specifically, S63 includes: matching the target noise variance level with a preset condition to obtain filtering parameters of the coil unit.

[0122] For example, the preset conditions are:

[0123] Condition 1: So w=3, iter=10.

[0124] Condition 2: So w=5, iter=15.

[0125] Condition 3: So w=7, iter=20.

[0126] Condition 4: So w=9, iter=25.

[0127] In this way, by adaptively matching the target noise variance level with the preset conditions, the filtering parameters of each coil unit can be obtained, so that the initial image of each coil unit can be weighted filtered based on the filtering parameters of each coil unit to obtain a filtered image.

[0128] In this embodiment, the initial noise variance level of the initial image is first determined based on the initial image of the coil unit, and the initial noise variance level is normalized. This can improve the speed and accuracy of obtaining the target noise variance, thereby more quickly determining the filter parameters of the coil unit based on the target noise variance level. This, in turn, improves the overall speed and accuracy of image reconstruction.

[0129] It should be understood that although Figure 2 、 5The steps in the flowcharts of FIG6 are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Figure 2 、 5 At least part of the steps in 6 may include multiple steps or multiple stages. These steps or stages do not necessarily have to be performed at the same time, but can be performed at different times. The order of execution of these steps or stages is not necessarily one by one, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0130] In one embodiment, referring to Figure 7 , provides an image reconstruction device 11, the device 11 includes: a data acquisition module 111, an initial image acquisition module 112, a weighted filtering processing module 113 and an image reconstruction module 114; specifically:

[0131] The data acquisition module 111 is used for multiple K-space data sets of the region of interest, each K-space data set corresponds to a coil unit in the magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0132] The initial image acquisition module 112 is configured to obtain an initial image corresponding to the coil unit according to the K-space data set acquired by the data acquisition module 111 .

[0133] The weighted filtering processing module 113 is configured to perform weighted filtering on the initial image obtained by the initial image acquisition module 112 to obtain a filtered image corresponding to the initial image.

[0134] The image reconstruction module 114 is used to obtain sensitivity estimates of each coil unit.

[0135] The image reconstruction module 114 is further configured to reconstruct the filtered image corresponding to each coil unit based on the sensitivity estimation of each coil unit obtained by the weighted filtering processing module 113 and the sensitivity parallel imaging algorithm to obtain a target image.

[0136] In one embodiment, the weighted filtering processing module 113 is specifically configured to determine filtering parameters of the coil unit and perform weighted filtering on the initial image corresponding to the coil unit based on the filtering parameters to obtain a filtered image corresponding to the initial image.

[0137] In one embodiment, the weighted filtering processing module 113 is specifically configured to determine an initial noise variance level of the initial image based on the initial image of the coil unit, normalize the initial noise variance level to obtain a target noise variance, and determine filtering parameters for the coil unit based on the target noise variance level.

[0138] In one embodiment, the weighted filtering processing module 113 is specifically configured to self-match the target noise variance level with a preset condition to obtain the window size and the number of iterations in the filtering parameters of the coil unit.

[0139] In one embodiment, the weighted filtering processing module 113 is specifically configured to calculate the weights of the filtering parameters of the coil units according to the target noise variance level.

[0140] In one embodiment, the weight is any one of a Gaussian weight and a reciprocal weight.

[0141] The specific definitions of the image reconstruction device can be found in the definitions of the image reconstruction method above and will not be repeated here. Each module in the aforementioned image reconstruction device may be implemented in whole or in part via software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor in a computer device in hardware form, or may be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0142] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store initial data, and the network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, an image reconstruction method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0143] Those skilled in the art will understand that Figure 8The structure shown is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0144] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0145] A plurality of K-space data sets of a region of interest are acquired, each K-space data set corresponding to a coil unit in a magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0146] An initial image corresponding to the coil unit is obtained according to the K-space data set.

[0147] Perform weighted filtering on the initial image to obtain a filtered image corresponding to the initial image.

[0148] Get the sensitivity estimate of each coil unit.

[0149] According to the sensitivity estimation of each coil unit and the sensitivity parallel imaging algorithm, image reconstruction is performed on the filtered image corresponding to each coil unit to obtain a target image.

[0150] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0151] A plurality of K-space data sets of a region of interest are acquired, each K-space data set corresponding to a coil unit in a magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0152] An initial image corresponding to the coil unit is obtained according to the K-space data set.

[0153] Perform weighted filtering on the initial image to obtain a filtered image corresponding to the initial image.

[0154] Get the sensitivity estimate of each coil unit.

[0155] According to the sensitivity estimation of each coil unit and the sensitivity parallel imaging algorithm, image reconstruction is performed on the filtered image corresponding to each coil unit to obtain a target image.

[0156] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0157] A plurality of K-space data sets of a region of interest are acquired, each K-space data set corresponding to a coil unit in a magnetic resonance coil; the K-space data set is obtained by undersampling the corresponding coil unit.

[0158] An initial image corresponding to the coil unit is obtained according to the K-space data set.

[0159] Perform weighted filtering on the initial image to obtain a filtered image corresponding to the initial image.

[0160] Get the sensitivity estimate of each coil unit.

[0161] According to the sensitivity estimation of each coil unit and the sensitivity parallel imaging algorithm, image reconstruction is performed on the filtered image corresponding to each coil unit to obtain a target image.

[0162] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0163] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0164] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. An image reconstruction method, characterized in that: include: Acquire multiple K-space data sets of a region of interest, each K-space data set corresponding to a coil unit in the magnetic resonance coil; The K-space data set is obtained by undersampling the corresponding coil unit; Obtaining an initial image corresponding to the coil unit according to the K-space data set; Performing weighted filtering on the initial image to obtain a filtered image corresponding to the initial image; the weighted filtering includes low-pass filtering; Obtaining sensitivity estimates for each coil unit; Reconstructing the filtered images corresponding to the coil units according to the sensitivity estimation of the coil units and the sensitivity parallel imaging algorithm to obtain a target image; The performing weighted filtering on the initial image to obtain a filtered image corresponding to the initial image includes: determining an initial noise variance level of the initial image according to an initial image of the coil unit; performing normalization processing on the initial noise variance level to obtain a target noise variance; and determining a filtering parameter of the coil unit according to the target noise variance level; Based on the filtering parameters, weighted filtering is performed on the initial image corresponding to the coil unit to obtain a filtered image corresponding to the initial image.

2. The image reconstruction method according to claim 1, wherein: The determining the filtering parameters of the coil unit according to the target noise variance level includes: The target noise variance level is matched with a preset condition to obtain a window size and a number of iterations in the filter parameters of the coil unit.

3. The image reconstruction method according to claim 1 or 2, characterized in that: The determining the filtering parameters of the coil unit according to the target noise variance level includes: The weights of the filter parameters of the coil unit are obtained by calculation according to the target noise variance level.

4. The image reconstruction method according to claim 3, wherein: The weight is either a Gaussian weight or a reciprocal weight.

5. The image reconstruction method according to claim 1, wherein: Obtaining a sensitivity estimate of each coil unit includes: For each of the coil units, a low-resolution volume coil image and a low-resolution surface coil image are collected; the volume coil image is divided by the surface coil image to obtain an initial sensitivity estimate of the coil unit; and Gaussian smoothing is performed on the initial sensitivity estimate to obtain a sensitivity estimate of the coil unit.

6. The image reconstruction method according to claim 1, wherein: The step of reconstructing the filtered images corresponding to the coil units according to the sensitivity estimation of the coil units and the sensitivity parallel imaging algorithm to obtain a target image includes: According to the sensitivity estimation of each coil unit and the sensitivity parallel imaging algorithm, overlapping parts of the filtered images corresponding to each coil unit are expanded and combined to obtain the target image.

7. An image reconstruction device, characterized in that: include: A data acquisition module is used to acquire multiple K-space data sets of the region of interest, each K-space data set corresponding to a coil unit in the magnetic resonance coil; The K-space data set is obtained by undersampling the corresponding coil unit; an initial image acquisition module, configured to obtain an initial image corresponding to the coil unit according to the K-space data set acquired by the data acquisition module; A weighted filtering processing module, configured to perform weighted filtering on the initial image obtained by the initial image acquisition module to obtain a filtered image corresponding to the initial image; the weighted filtering processing includes low-pass filtering processing; An image reconstruction module for obtaining sensitivity estimates of each coil unit; The image reconstruction module is further configured to reconstruct the filtered image corresponding to each coil unit according to the sensitivity estimation of each coil unit obtained by the weighted filtering processing module and the sensitivity parallel imaging algorithm to obtain a target image; The weighted filtering processing module is specifically used to determine the initial noise variance level of the initial image based on the initial image of the coil unit; normalize the initial noise variance level to obtain a target noise variance; determine the filtering parameters of the coil unit based on the target noise variance level; and perform weighted filtering on the initial image corresponding to the coil unit based on the filtering parameters to obtain a filtered image corresponding to the initial image.

8. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the computer program is executed by the processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images