Magnetic resonance imaging apparatus and magnetic resonance image generation method

Abstract

This improves the accuracy of motion-corrected reconstruction in response to body movements that occur during MRI imaging, thereby improving the accuracy of motion-corrected images. [Solution] When collecting k-space data consisting of nuclear magnetic resonance signals with an MRI device, the intensity of motion correction is varied according to the k-space position of the measurement data affected by the generated body movement. In one embodiment, when determining whether the measurement data acquired during motion detection is affected by body movement, the threshold value used for motion determination is varied according to the k-space position of the measurement data. In another embodiment, when performing motion correction reconstruction including iterative calculations, motion-affected data is not excluded as the initial value of the iterative calculations, but when incorporating it into the iterative calculations or during the processing of the iterative calculations, the processing weights are varied according to the k-space position of the motion-affected data to perform motion correction reconstruction.

Description

Technical Field

[0001] The present invention relates to a magnetic resonance imaging apparatus (hereinafter referred to as an MRI apparatus), and particularly to a technique for reducing the influence of body movement generated in a subject during imaging.

Background Art

[0002] MRI apparatuses can non-invasively grasp the internal tissues of a subject to be examined and are widely used in the medical field. One problem for obtaining good images with an MRI apparatus is the body movement of the subject during imaging. The body movement of the subject includes movements associated with breathing and heartbeat and other unintentional movements, for example, movements associated with physiological reactions such as spasms, sneezes, and coughs, and other sudden movements. Such body movement generates artifacts in the image, and the resulting image quality degradation hinders image diagnosis.

[0003] Conventionally, various techniques for reducing the degradation of image quality due to body movement have been proposed. In dealing with body movement, detection of the body movement of the subject during imaging and processing of the data collected at the time when the body movement is detected are important processes. In an MRI apparatus, detection of body movement that occurs while collecting k-space data is a problem. In parallel with the collection of k-space data for image reconstruction, a nuclear magnetic resonance signal (referred to as a navigator echo) for detecting body movement is collected, and a method of detecting body movement from the navigator echo (for example, Patent Document 1), a method of installing instruments such as a pressure sensor and an abdominal pressure gauge to mainly detect respiratory movement, and a method of installing a monitoring camera in the imaging space where the subject is placed and analyzing the frame image from the camera to detect body movement (Patent Document 2) have been proposed. Note that Patent Document 1 discloses dividing the k-space into two or more blocks and detecting body movement by different body movement detection means (one of which is a navigator echo) for each block.

[0004] Processing of data collected when body movement is detected (hereinafter referred to as body movement effect data) can be done using phase correction of the body movement effect data using phase information obtained from navigator echo (Patent Document 1), or by replacing the signal of the body movement occurrence location with zero and reconstructing it, or by estimating and reconstructing the body movement effect data through iterative calculations, etc. (Patent Document 3, etc.). [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 06-047021 [Patent Document 2] Japanese Patent Publication No. 2023-027608 [Patent Document 3] Japanese Patent Publication No. 2023-022669 [Overview of the project] [Problems that the invention aims to solve]

[0006] Conventional motion processing methods sometimes take into account differences in motion levels depending on the region or part of the subject. However, the impact of motion on an image can be influenced by various factors other than the motion level, and conventional methods may over-remove measurement data that has a relatively small impact on the image, or fail to adequately remove measurement data that has a large impact on the image. In such cases, problems such as insufficient removal of motion artifacts or blurred images may occur.

[0007] For example, Patent Document 1 discloses a method in which image data is divided into regions with relatively large and relatively small motion magnitudes, each region is converted into measurement spatial data, and different phase corrections are performed on the measurement data for each region. However, although this method allows for motion correction that takes into account the magnitude of motion, it cannot solve the aforementioned problems.

[0008] The present invention aims to improve the accuracy of motion processing by performing processing that takes into account the region of k-space. [Means for solving the problem]

[0009] To solve the above problems, the present invention provides a mechanism for collecting k-space data consisting of nuclear magnetic resonance signals using an MRI device, which varies the intensity of motion correction depending on the position of the measurement data in k-space that is affected by the generated body movement. In one embodiment, when determining whether the measurement data acquired during motion detection is affected by body movement, the threshold value used for motion determination is varied depending on the position of the measurement data in k-space. Specifically, the threshold is set so that motion determination is stricter in the low-frequency range of k-space and lenient in the high-frequency range.

[0010] In another embodiment, when performing motion correction reconstruction including iterative calculations, motion-corrected reconstruction is performed in which, instead of excluding motion-affected data as initial values ​​for the iterative calculations, the processing weights are varied according to the k-space position of the motion-affected data when incorporating it into the iterative calculations or during the processing of the iterative calculations, thereby reducing the effect of motion.

[0011] In other words, the MRI apparatus according to the first aspect of the present invention comprises an imaging unit that collects k-space data consisting of magnetic resonance signals, and a processor that generates an image using the k-space data and analyzes the body movement of a subject during imaging. The processor detects the body movement of a subject during imaging using information from an optical imaging device that optically photographs the subject during imaging or magnetic resonance signals obtained by the imaging unit, identifies data affected by body movement among the k-space data as body movement affected data, and is characterized in that the threshold used to determine whether or not a data is body movement affected data is different depending on the position of the data to be determined in k-space.

[0012] Furthermore, the MRI apparatus according to the second aspect of the present invention is characterized in that the processor detects the movement of the subject during imaging using information from an optical imaging device that optically photographs the subject during imaging or a magnetic resonance signal obtained by the imaging unit, identifies data affected by the movement among the k-space data as movement-affected data, and performs a movement correction reconstruction including iterative calculations using the k-space data including the movement-affected data, and in this movement correction reconstruction, processes are performed to reduce the effect of the movement on the movement-affected data.

[0013] Furthermore, the present invention provides a method for generating motion-corrected images using k-space data acquired by a magnetic resonance imaging apparatus, which includes body motion information at the time of acquisition as supplementary information.

[0014] A magnetic resonance image generation method according to a first aspect of the present invention is characterized by using motion information to identify motion-affected data among k-space data that is affected by motion, and by making the threshold value for motion used to determine the motion-affected data different depending on the position in k-space.

[0015] A second aspect of the present invention provides a magnetic resonance image generation method characterized by identifying data affected by motion from k-space data as motion-affected data, performing motion correction reconstruction including iterative calculations using k-space data including motion-affected data, and then using an image obtained by converting the k-space data including the smoothed motion-affected data into an image space as the initial value for the iterative calculations. [Effects of the Invention]

[0016] According to a first aspect of the present invention, by changing the threshold for motion detection depending on whether the measurement data collected at the time of motion occurrence is low-frequency or high-frequency data in k-space, it is possible to prevent insufficient or excessive removal of motion-affected data, prevent inadequate removal of motion artifacts or, conversely, blurring of the image due to excessive removal, and enable highly accurate motion correction.

[0017] According to the second aspect of the present invention, a process for reducing the weight of body movement influence data can be included in the body movement correction reconstruction including iterative calculations, thereby performing effective body movement removal while reducing the cost of iterative calculations. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] [Figure 1] A diagram showing an overall overview of an MRI apparatus to which the present invention is applied [Figure 2] A flowchart showing an overview of the processing of a processor [Figure 3] A flowchart showing the processing flow of Embodiment 1 [Figure 4] A diagram for explaining a region of k-space [Figure 5] A graph showing the relationship between positions on k-space and a threshold value [Figure 6] A diagram showing the relationship between body movement information and a pulse sequence [Figure 7] A diagram showing an example of body movement correction reconstruction by a processor (image generation unit) of Embodiment 1 [Figure 8] A diagram for explaining smoothing of body movement influence data [Figure 9] A diagram showing the processing flow of a modification of Embodiment 1 DETAILED DESCRIPTION OF THE INVENTION

[0019] Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0020] First, an overview of the MRI apparatus to which the present invention is applied will be described. As shown in Figure 1, the MRI apparatus 1 includes an imaging unit 10 that generates nuclear magnetic resonance in the atomic nuclei of atoms constituting the tissue of a subject 50 and collects nuclear magnetic resonance signals generated from the subject; a processor 20 that processes the nuclear magnetic resonance signals collected by the imaging unit 10 and controls the imaging unit 10; and a user interface unit (hereinafter referred to as UI unit) 30 for the MRI apparatus 1 and an operator such as a doctor or technician (hereinafter referred to as user) to set imaging conditions, input commands necessary for processing, and display images and GUIs obtained by the MRI apparatus 1. The MRI apparatus 1 may also be equipped with an external storage device 60 for storing generated images and other information, and an interface (not shown) for communicating with external devices.

[0021] The configuration of the imaging unit 10 is similar to that of a known MRI device, and includes a static magnetic field magnet 101 that generates a uniform static magnetic field in the examination space where the subject 50 is placed, a gradient magnetic field coil 102 that gives a magnetic field gradient to the static magnetic field, an RF transmitting coil 103 that applies a predetermined high-frequency magnetic field to the subject, and an RF receiving coil 104 that receives a nuclear magnetic resonance signal (hereinafter also referred to as an echo signal) generated from the subject. The gradient magnetic field coil 102 is connected to a gradient magnetic field power supply 105, the RF transmitting coil 103 is connected to an RF transmitting unit 106 consisting of an RF oscillator and an RF amplifier, and the RF receiving coil 104 is connected to an RF receiving unit 107 equipped with a QD detector and an AD converter. While a single RF coil may serve both purposes, generally, the RF transmitting coil 103 is housed within a gantry (not shown) that surrounds the inspection space, along with the static magnetic field magnet 101 and the gradient magnetic field coil 102, while the RF receiving coil 104 is placed in the inspection space attached to the subject 50.

[0022] The imaging unit 10 further includes a sequencer 108 that operates the RF transmission unit 106, gradient magnetic field power supply 105, and RF reception unit 107 according to a predetermined pulse sequence, and imaging is performed according to the imaging sequence set in the sequencer 108. The operations related to imaging are the same as those of a general MRI device, and in this specification, a description is omitted except for the processes related to the invention.

[0023] The imaging unit 10 is further equipped with a bed device 40 for placing the subject. In addition, one or more surveillance cameras 80 for monitoring the subject may be installed inside the gantry or at the end of the opening. In that case, the surveillance cameras 80 can also function as motion detection means for detecting the subject's body movements.

[0024] The processor 20 controls the imaging unit 10, performs signal processing and various calculations on the nuclear magnetic resonance signals collected by the imaging unit 10, and performs processing (motion processing) to reduce the effects of body movements that occur in the subject during imaging, as described above. For example, the processor 20 takes in video from the surveillance camera 80, monitors for the occurrence of body movements, and, based on the information at the time of the movement, applies a timestamp to the body movement in the measurement data collected at that time and performs processing such as motion correction.

[0025] To achieve the above processing, the processor 20 includes an imaging control unit 210 that controls imaging, a display control unit 250 that controls the display in the UI unit 30, an image generation unit 220 that performs various calculations related to image generation such as image reconstruction, and a body motion processing unit 230 that processes the body movements of the subject 50 during imaging.

[0026] Although the processor 20 is shown as a single block in Figure 1, it may be composed of one or more hardware components or a combination of hardware and a program. The type of hardware is not limited; for example, the processor may be composed of a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a programmable logic device such as an FPGA (Field Programmable Gate Array), a dedicated circuit for executing a specific process such as an ASIC (Application Specific Integrated Circuit), a GPU (Graphic Processing Unit), or an NPU (Neural Processing Unit). Furthermore, the type of hardware may be a combination of different types of hardware. When multiple hardware components are configured to execute one or more processes of a given processor, these multiple hardware components may reside in physically separate devices or in the same device.

[0027] When the processor 20 is implemented as a combination of hardware and a program, the program may be firmware or software such as microcode. Alternatively, the program may be, for example, a group of program modules, and each of its functions may be implemented by a processor configured to perform its respective function. The program may also be program code or multiple code segments stored in one or more non-temporary computer-readable media (e.g., storage media or other storage).

[0028] Based on the above configuration, the processing overview of the MRI apparatus 1 of this embodiment will be explained. Figure 2 shows the processing flow.

[0029] First, when imaging is started under the control of the imaging control unit 210, the motion processing unit 230 acquires motion information detected by motion detection means such as the surveillance camera 80 (S1), and determines whether motion affecting the image has occurred using a threshold (S2). Whether motion affects the image depends on the level and duration of the motion and the position in k-space of the measurement data acquired at the time of motion. In this embodiment, the motion level is determined using different thresholds depending on the position in k-space of the measurement data, and it is determined whether the measurement data affects the image (motion-affected data).

[0030] The threshold for determining body movement can be set in advance by pre-determining a standard body movement level threshold for each body part, or by determining a standard body movement level threshold based on the body movement levels for each patient or each body part of the patient obtained in preliminary measurements, and then adjusting the standard threshold according to the position in k-space when applying the body movement determination. Alternatively, the standard threshold can be set in advance according to the position in k-space, stored in memory as, for example, a table or function, and read out and used when applying the determination.

[0031] When motion-affect data is identified during motion detection, the subsequent processing is determined based on the position of the motion-affect data in k-space and the number of motion-affect data (S3). Subsequent processing includes motion correction reconstruction, such as remeasurement, removal of motion-affect data, correction, and zero substitution, and the system determines which processing to perform. These processing methods can be selected based on predetermined criteria.

[0032] Next, imaging or image reconstruction is performed according to the determined process, and finally, once the measurement data necessary for image reconstruction has been collected, the image generation unit 220 performs image reconstruction (S4). Image reconstruction includes image reconstruction with correction for body motion (body motion correction reconstruction) if one or more body motion effect data are identified. When performing body motion correction reconstruction, the image generation unit 220 may employ a method that takes into account the position of the body motion effect data in k space. For example, when performing iterative calculations as body motion correction reconstruction, the weights used in the calculation may be different depending on the position of the measurement data in k space. The image generated by the image generation unit 220 is displayed on the display device of the UI unit 30 by the display control unit 250, along with supplementary information such as subject information and imaging conditions (S5). It is also transferred to an external storage device or an external device other than the MRI device 1 as needed.

[0033] According to the MRI apparatus of this embodiment, the accuracy of identifying motion-affected data can be improved by adjusting the threshold used for motion detection according to the position of the measurement data in k-space measured when motion is detected. This prevents over- or under-correction of motion and enables the provision of high-quality motion-corrected images with good time efficiency.

[0034] The following describes specific embodiments of the MRI apparatus processing of this embodiment, focusing on threshold setting considering k-space position and motion correction reconstruction considering k-space position.

[0035] <Embodiment 1> In this embodiment, we describe a case in which, in setting the threshold, adjustments are made according to the k-space position of the measurement data, and reconstruction is performed by iterative calculation as a motion correction reconstruction method, and in that case, weighting according to the k-space position is performed in the data consistency processing.

[0036] <<Example of threshold setting>> Figure 3 shows the processing flow of the embodiment related to threshold setting.

[0037] First, a reference threshold is set in advance (S31). The threshold is determined by the magnitude of the motion level that affects the image, and the degree to which the motion level affects the image varies depending on the imaging area and the image resolution. Therefore, as a simple method, a predetermined threshold can be set in advance for each area according to the planned image quality. Alternatively, the motion characteristics of each subject can be obtained in pre-measurements (such as imaging to determine the imaging cross-section and imaging conditions) performed before the main imaging (imaging to acquire an image of the subject), and a threshold for each subject, and further, if necessary, a threshold for each part of the subject, can be set based on these characteristics.

[0038] Based on the threshold set in this way, thresholds corresponding to the k-space position are set. The setting of thresholds according to the k-space position is basically based on the idea that body movement detection should be stricter in the low-frequency range of k-space and lenient in the high-frequency range. Specifically, for example, k-space is divided into a low-frequency range and a high-frequency range, with a smaller threshold in the low-frequency range and a larger threshold in the high-frequency range. Figure 4 shows examples of k-space division. Division example 1 shown on the left side of Figure 4 divides the region above and below the k-space origin (0-encoded) into a low-frequency range and a high-frequency range, and sets the thresholds for the low-frequency range and the high-frequency range to different values ​​(low-frequency threshold < high-frequency threshold). Division example 2 shown on the right side of Figure 4 further divides the low-frequency range and the high-frequency range, and sets the thresholds to increase in the order of low-frequency 2, low-frequency 1, high-frequency 2, and high-frequency 1.

[0039] There are no particular limitations on the degree to which the thresholds differ between low and high frequencies. However, even with high-frequency measurement data, if the magnitude of body movement is too large, temporally adjacent measurement data may also be affected by large body movements. Therefore, if the low-frequency threshold is set to the predetermined standard threshold as described above, the high-frequency threshold should be set to, for example, 1.5 to 2 times the low-frequency threshold. Conversely, if the high-frequency threshold is set to the predetermined standard threshold, the low-frequency threshold should be set to 0.8 to 0.5 times the high-frequency threshold. These ratios can be predetermined for each divided region and applied based on the position in k-space during the threshold-based judgment process.

[0040] Alternatively, instead of dividing the k-space and applying thresholds, it is also possible to continuously change the threshold from low to high frequencies, as shown in Figure 5. Figure 5 shows an example where the threshold is changed linearly (by a linear function) for phase encoding, but a quadratic function or other function may also be used. In these cases, the function may be stored in memory and read out during the threshold-based determination process.

[0041] The ratio by which the threshold for detecting body movement differs between low and high frequencies, or the relationship with the threshold for phase encoding, may be set in advance, or it may be set by the user via the UI unit 30 and used by the body movement processing unit 230.

[0042] Next, when the main imaging begins, motion information of the subject is acquired in parallel (S32). There are two main methods for acquiring motion information: one is to acquire it by analyzing frame images sent from the surveillance camera 80, and the other is to acquire it by analyzing nuclear magnetic resonance signals generated from the subject, such as navigator echoes. Either method may be used. The individual methods are well known, and detailed explanations are omitted here, but for example, in the analysis of frame images, motion vectors for each pixel between frame images are calculated using methods such as optical flow, and the amount of motion is determined. In the analysis of navigator echoes, in addition to the nuclear magnetic resonance signals collected in the main imaging, there are methods such as generating and collecting nuclear magnetic resonance signals (navigator echoes) without applying phase encoding, and acquiring positional changes in real space from the data obtained by Fourier transforming it in one dimension, or analyzing from changes in the navigator echoes themselves. Either method may be used.

[0043] When motion is detected, the k-space position of the measurement data collected at the time of motion detection is obtained (S33). Since the k-space position of the measurement data is determined by the phase encoding assigned to that measurement data, the motion processing unit 230 can obtain the k-space position from the progress of the pulse sequence. Figure 6 shows the relationship between motion information and pulse sequence. The example in Figure 6 shows the case where the motion processing unit 230 obtains motion information from frame images sent from the surveillance camera 80, and the pulse sequence is a spin echo (SE) system sequence.

[0044] Next, the position of the measurement data in k-space obtained in step S33 is referenced, and it is determined whether the generated body movement affects the image using a threshold corresponding to the position in k-space (S34). If the detected body movement exceeds the threshold as a result of the determination, the measurement data at the time the body movement occurred is labeled as body movement effect data, as shown in Figure 6.

[0045] The motion processing unit 230 determines the processing after motion detection and during image reconstruction according to the continuity of motion, i.e., the continuity of the identified motion effect data, the position or proportion of the motion effect data in k-space, imaging conditions, and image reconstruction conditions (S35). For example, if a large amount of motion effect data occurs in the low region of k-space and has a significant impact on the image, the motion effect data may be recaptured. In that case, the determination result is passed to the imaging control unit 210, which controls the pulse sequence to collect the identified motion effect data and subsequent unmeasured data. If the number of motion effect data is small and can be estimated by calculation, imaging is continued as is, and a processing determination is made to perform a predetermined motion correction reconstruction during image reconstruction. There are various methods for determining the processing, and the method is not limited to the method described above. For example, it is possible to use various known methods, such as adding a more detailed determination flow as disclosed in Patent Document 3, or using criteria other than the position or number in k-space.

[0046] When motion correction reconstruction is performed, the process is the same as in Figure 3, and motion correction reconstruction (S36) and display of the reconstructed image (S37) are performed.

[0047] When motion-affect data is identified, there are several known methods for motion correction reconstruction, such as removing the motion-affect data and replacing it with zeros, correcting the phase of the motion-affect data based on the magnitude of the motion, estimating the motion-affect data from other measurement data using Hermitian symmetry in k-space or by applying the PI (Parallel Imaging) method and reconstructing the image, or iterative reconstruction in which the motion-affect data is removed and estimated through iterative calculations as unmeasured data. Any of these methods can be adopted.

[0048] As explained above, in this embodiment, in the body motion detection, the threshold for body motion detection is varied depending on the position in k-space of the measurement data collected when body motion occurs. This reduces the excess or deficiency of body motion effect data to be identified, improving the accuracy of identifying body motion effect data and subsequently improving the accuracy of body motion correction reconstruction.

[0049] <<Embodiment of Motion Correction Reconstruction>> The following describes an embodiment of the process when iterative calculations are used as the motion correction reconstruction.

[0050] As shown in Figure 7, an example of motion correction using iterative calculations is performed by using measurement data 701 (its image space data 710), which is obtained by removing motion-affect data identified by motion detection from measurement data 700, as the initial value. This is followed by repeated estimation interpolation of unmeasured data (data that was removed and is missing), conversion of the estimated measurement data 712 to image space data, and updating of the initial value. During this iterative calculation, an integration process (hereinafter also called data consistency processing) is performed to improve the degree of agreement between the estimated measurement data 712 and the original measurement data 700, and the final motion-corrected image 720 is generated.

[0051] In this embodiment, while conventional methods removed motion effect data from the measurement data 700, this embodiment applies a smoothing filter to the motion effect data and uses measurement data that includes the smoothed motion effect data. This measurement data including the smoothed motion effect data is inversely Fourier transformed into real-space data, and the resulting image is used as the initial value (initial image) for iterative calculations. Here, the smoothing filter is a one-dimensional filter in the frequency direction for k-space data, and a blurring filter such as a Gaussian filter can be used. In this case, the intensity of the Gaussian filter may be varied according to the motion level. That is, the filter intensity is increased when the motion level is high, and the standard filter intensity is used when the motion level is low (for example, close to the threshold).

[0052] When multiple motion effect data exist, either a method for smoothing each individual motion effect data or a method for additively smoothing multiple motion effect data can be employed. An example of the former is shown in the upper part of Figure 8, and an example of the latter is shown in the lower part of Figure 8. Here, as an example, we assume that three motion effect data have been identified within the k-space data. As shown in the upper part of Figure 8, when smoothing individual motion effect data, the one-dimensional filter described above is applied to each of the three motion effect data. In this case, the filter strength may be varied according to the distance of the motion effect data from the center of k-space (phase encoding 0). That is, the filter strength may be stronger the closer to the center and weaker the further away from the center.

[0053] Furthermore, as shown in the lower part of Figure 8, when summarizing and smoothing multiple motion effect data, first, the summation average of the multiple motion effect data is taken, a one-dimensional filter is applied to the summed data, and the filtered data is rearranged to the positions of the original motion effect data. When rearranging the summed data, the signal intensity may be weighted according to the distance from the center of the k-space of the motion effect data (phase-encoded 0). Specifically, when rearranging the data after additive averaging smoothing at the location of the low-frequency motion-affect data, the signal intensity is weighted to be higher to match the surrounding low-frequency data. When rearranging the data at the location of the high-frequency motion-affect data, the signal intensity is weighted to be lower to match the surrounding high-frequency data.

[0054] Furthermore, when averaging, each motion effect data may be assigned different weights depending on its distance from the center of k-space. For example, high-frequency measurement data may be given a smaller weight, and low-frequency measurement data a larger weight. Alternatively, the weights may be varied according to the motion level of the motion effect data. That is, when the motion level acquired during motion detection is high, the weight of the measurement data may be small, and when the motion level is low, the weight of the measurement data may be relatively large.

[0055] As described above, by filtering out the motion-affect data that was previously removed and using it for motion correction reconstruction, the initial image obtained by converting the filtered motion-affect data still contains the effects of motion, but it also contains more information from the original measurement data, thus reducing the computational cost of the iterative motion correction reconstruction calculations.

[0056] Furthermore, by applying filtering, it is possible to avoid re-shooting motion-affect data occurring in the low frequency range, thus preventing the prolonged imaging time associated with re-shooting. Furthermore, when summarizing and smoothing multiple motion effect data, a greater reduction in computational cost can be achieved by weighting the motion effect data according to the magnitude of its influence on motion, i.e., weighting it according to the position in k-space or the motion level.

[0057] The two methods described above perform filtering on motion effect data, but preliminary iterative calculations may also be performed to estimate the data as processing for motion effect data. For example, iterative calculations for motion correction are performed using only motion effect data as the initial value. The k-space data obtained from these iterative calculations (data corresponding to each motion effect data) are averaged to return to the original measurement data, and this is converted to image data and used as the initial value for the iterative calculations. In this method as well, by averaging, the effect of leveling out positional displacement due to motion is obtained, similar to the case where the data after the averaging smoothing process described above is rearranged, and it has the advantage of being able to use the motion effect data without removing it. However, in order to suppress the increase in computational cost due to performing preliminary iterative calculations, it is preferable to apply high-intensity denoising to the preliminary iterative calculations and complete the convergence quickly.

[0058] In the motion-corrected reconstruction of this embodiment, iterative calculations are performed using the initial image obtained after processing the motion-affect data as described above. This makes it possible to quickly obtain motion-corrected estimated k-space data while making use of the information from the motion-affect data.

[0059] Motion correction reconstruction is performed in the manner shown in Figure 7 as an example, but in addition to using the initial image processed as described above, modifications that take k-space into account may be made in the iterative calculation itself. However, this modification is independent of the modification of the initial image due to the processing of the motion effect data itself as described above, and it is not necessary to make both modifications. In motion correction reconstruction, adopting either one of the modifications is also included in the motion correction reconstruction of this embodiment.

[0060] The iterative calculation repeatedly performs processes such as transforming the initial image 710 into k-space data (measurement data 701), minimizing the L2 norm, and performing data estimation to obtain estimated k-space data (estimated k-space data 712), and then updating the initial image 710. During this process, data consistency is maintained by integrating the estimated k-space data 712 (which is still being processed) and the k-space data obtained by transforming the updated initial image (k-space data before estimation 711) with the original measurement data 700.

[0061] In this embodiment, during the data consistency processing, the k-space data 712 to be integrated is weighted according to the location in k-space where motion-affect data existed. The weights reflect at least one of the presence or absence of motion and the magnitude of the motion, and the weight of data at locations where motion occurred or where the motion was large is increased to enhance the strength of the correction. That is, the weight of data at locations in k-space where motion-affect data has been identified is made greater than the weight of data at locations in k-space where no motion-affect data existed. By performing such weighting, the strength of the correction of the effect of motion in the estimated k-space data 712 can be increased, and a final motion-corrected image 720 with further reduced motion can be obtained without image degradation or a decrease in data consistency.

[0062] Alternatively, instead of weighting the body movement effect data based on its position in k-space, or in addition to that, weighting may be performed considering the magnitude of the body movement itself. In that case, the body movement effect data is weighted according to the magnitude of the body movement at the time each body movement effect data was acquired, and the k-space data is weighted accordingly, with data that are greatly affected by body movement having a larger weight, and data that are little or no affected by body movement having a smaller weight.

[0063] In data consistency processing, by applying weights to the k-space data 712 as described above (weights according to the position of the motion-affected data in k-space, and weights according to the magnitude of the motion), the effect of removing the effects of motion, especially large motions, is strongly activated within a single iteration of the loop. Therefore, even with a limited number of iterations, the effect of reducing the effects of motion can be achieved.

[0064] As explained above, according to this embodiment, in the motion processing, by making the threshold for motion detection different depending on the position in k-space of the nuclear magnetic resonance signal (measurement data) acquired when motion occurs, it is possible to reduce over- or under-detection of motion and perform accurate motion detection.

[0065] Furthermore, according to this embodiment, when performing motion correction reconstruction including iterative calculations, the image is reconstructed by performing processing to reduce the effect of motion while making the most of the motion effect data. Processing to reduce the effect of motion includes, for example, smoothing the motion effect data and using it as the initial value for iterative calculations, and weighting the intermediate data of iterative calculations including the motion effect in data consistency processing according to the position in k-space of the data collected when the motion occurred and the magnitude of the motion. By performing such processing, the computational cost of motion correction reconstruction can be reduced.

[0066] Furthermore, the embodiment of threshold setting for body motion detection and the embodiment related to processing body motion effect data in body motion correction reconstruction in this embodiment are independent of each other, and combinations of the threshold setting embodiment with other body motion correction reconstruction methods, and combinations of other threshold setting methods with embodiments related to body motion correction reconstruction are also included in the present invention.

[0067] <Variation> Furthermore, although the above embodiments described a case in which imaging and subsequent image reconstruction are performed as a series of processes, it is also possible to identify motion effect data based on motion information (magnitude of motion and time of motion occurrence) attached to the measurement data after imaging and perform motion correction reconstruction separately from the image reconstruction performed as part of imaging, and the present invention also includes cases in which motion processing and / or motion correction reconstruction are performed retrospectively in this manner. In such retrospective processing, the setting of the threshold for motion determination and the setting of weights in motion correction, as described in the embodiments above, may be configured to be arbitrarily set by the user.

[0068] Figure 9 shows the processing flow when post-processing is performed. In Figure 9, processes that are the same as those in Figure 3 are indicated by the same reference numerals as in Figure 3, and redundant explanations are omitted. When performing post-processing image reconstruction, first k-space data with attached body motion information is acquired (S91). In this k-space data, the fact that it was acquired when body motion occurred and the body motion level at that time are labeled as attached information. When the user sets a desired threshold (S92), the body motion processing unit 230 identifies body motion effect data using that threshold (S34). The threshold set by the user may be a single threshold, but it is also possible to set thresholds according to the position in k-space, for example, different thresholds for low and high frequencies. The method is the same as the threshold embodiment described above. After that, predetermined processing, such as filtering, is applied to the identified body motion effect data, body motion correction reconstruction including iterative calculations is performed, and the generated image is displayed (S36, S37). The user can change the way the low and high frequency thresholds are differentiated as needed to perform body motion correction reconstruction (S93).

[0069] Although Figure 9 shows a case where only the threshold is user-defined, it is also possible to configure the system to allow user-defined conditions for motion correction reconstruction, such as the selection of the filtering method and the weights in data consistency.

[0070] According to this modification, it is possible to generate various motion-corrected images with different motion processing conditions, and the user can determine the optimal motion processing conditions. [Explanation of Symbols]

[0071] 1: MRI device, 10: imaging unit, 20: processor, 30: UI unit, 50: subject, 230: motion processing unit

Claims

1. An imaging unit that collects k-space data consisting of magnetic resonance signals, The system includes a processor that generates an image using the k-space data and analyzes the body movement of the subject during imaging, The processor detects the movement of the subject during imaging using information from an optical imaging device that optically photographs the subject during imaging or a magnetic resonance signal obtained by the imaging unit. A magnetic resonance imaging apparatus characterized by identifying data affected by body motion from the k-space data as motion-affected data, and by using a threshold value different for determining whether or not a data is motion-affected depending on its position in k-space.

2. A magnetic resonance imaging apparatus according to claim 1, The magnetic resonance imaging apparatus is characterized in that the processor sets the threshold to be low on the low-frequency side of k-space and high on the high-frequency side.

3. A magnetic resonance imaging apparatus according to claim 2, The magnetic resonance imaging apparatus is characterized in that the processor divides the k-space into a plurality of regions including low-frequency and high-frequency regions, and sets the threshold applied to the low-frequency data to be lower than the threshold applied to the high-frequency data.

4. A magnetic resonance imaging apparatus according to claim 1, The magnetic resonance imaging apparatus is characterized in that the processor generates a motion-corrected image by iterative calculation using k-space data including motion effect data.

5. A magnetic resonance imaging apparatus according to claim 4, The aforementioned processor is characterized by replacing motion effect data with zeros and generating a motion-corrected image, and is a magnetic resonance imaging apparatus.

6. A magnetic resonance imaging apparatus according to claim 4, The magnetic resonance imaging apparatus is characterized in that the processor performs a smoothing process on motion-affected data and generates a motion-corrected image using k-space data that includes the smoothed motion-affected data.

7. A magnetic resonance imaging apparatus according to claim 4, The magnetic resonance imaging apparatus is characterized in that, when k-space data includes multiple motion-affect data, the processor performs a smoothing process after adding the multiple motion-affect data, rearranges the smoothed motion-affect data to the original positions of the multiple motion-affect data in k-space, and then generates a motion-corrected image.

8. A magnetic resonance imaging apparatus according to claim 4, The aforementioned iterative calculation includes the conversion from k-space data to image-space data, data estimation processing in image space, conversion from estimated image-space data to k-space data, integration processing (data consistency processing) to ensure data consistency between the k-space data during the calculation and the measured data, and conversion of the k-space data after integration processing to image-space data. A magnetic resonance imaging apparatus characterized in that the weights of the k-space data during the calculation process in the integration process are varied according to the position of the motion-affected data in k-space.

9. An imaging unit that collects k-space data consisting of magnetic resonance signals, The system includes a processor that generates an image using the k-space data and analyzes the body movement of the subject during imaging, The processor detects the movement of the subject during imaging using information from an optical imaging device that optically photographs the subject during imaging or a magnetic resonance signal obtained by the imaging unit. The data affected by body movement from the aforementioned k-space data is identified as body movement-affected data, and a body movement correction reconstruction including iterative calculations is performed using the k-space data including the aforementioned body movement-affected data. A magnetic resonance imaging apparatus characterized in that, in the motion correction reconstruction described above, processing is performed to reduce the effect of motion on the motion effect data.

10. A method for generating motion-corrected images using k-space data acquired by a magnetic resonance imaging apparatus, the k-space data including motion information at the time of acquisition as supplementary information, A magnetic resonance image generation method characterized by using the aforementioned body motion information to identify body motion-affected data from the k-space data, and by making the threshold value for body motion used to determine the body motion-affected data different depending on the position in k-space.

11. A method for generating magnetic resonance images according to claim 10, The step includes accepting the setting or modification of the threshold by the user, A magnetic resonance image generation method characterized by generating motion-corrected images each time the threshold differs due to settings or changes.

12. A method for generating motion-corrected images using k-space data acquired by a magnetic resonance imaging apparatus, the k-space data including motion information at the time of acquisition as supplementary information, Data affected by body movement within the k-space data is identified as body movement-affected data, and body movement correction reconstruction, including iterative calculations, is performed using the k-space data including the body movement-affected data. A magnetic resonance image generation method characterized by performing a smoothing process on the motion effect data, and then using an image obtained by converting the k-space data, which includes the smoothed motion effect data, into an image space as the initial value for iterative calculations.

13. A magnetic resonance image generation method according to claim 10 or 12, The aforementioned iterative calculation includes the conversion from k-space data to image-space data, data estimation processing in image space, conversion from estimated image-space data to k-space data, integration processing (data consistency processing) for data consistency between k-space data during the calculation and measured data, and conversion of the integrated k-space data to image-space data. A method for generating magnetic resonance images, characterized in that the weights of the k-space data during the calculation process in the aforementioned integration process are made different according to the position of the motion-affected data in k-space.

Images