JP4721857B2 - Magnetic resonance imaging apparatus and image reconstruction method using the same - Google Patents

Description

The present invention relates to a magnetic resonance imaging (hereinafter referred to as “MRI”) apparatus that obtains a tomographic image of a desired region of a subject using a nuclear magnetic resonance (hereinafter referred to as “NMR”) phenomenon, and particularly using a plurality of receiving coils. The present invention relates to a reconstruction processing method performed on an acquired NMR signal.

An MRI apparatus detects a NMR signal generated from a subject by exciting a nucleus such as hydrogen contained in the subject by irradiating a subject placed in a static magnetic field with a high frequency magnetic field of a specific frequency. The physical or chemical information of the subject can be acquired and imaged.
In such an MRI apparatus, a high-frequency coil is used to detect an NMR signal. As this high-frequency coil, a coil composed of a plurality of receiving coils called a phased array coil or a multiple coil is used because of its good sensitivity and a good SN ratio even when the field of view is widened.

  In such a multiple coil, NMR signals obtained from a plurality of receiving coils constituting the image are respectively reconstructed, and then the reconstructed images are synthesized and displayed as a final image. For this purpose, each receiving coil has a preamplifier and an A / D converter, and each NMR signal amplified by this preamplifier and digitized by the A / D converter has a number of transmission paths (channels) equal to or greater than the number of receiving coils. Is input to a signal processing system. In this signal processing system, an arithmetic unit for processing the NMR signal is provided for each transmission path, and the NMR signal input to the transmission path is processed. That is, a plurality of arithmetic units are arranged in parallel in the signal processing system, and a large amount of data is processed at high speed by processing a plurality of NMR signals in parallel.

As an example of such a configuration, for example, (Patent Document 1) includes A / D conversion means for digitizing an NMR signal and data for reconstructing an image by performing arithmetic processing on the digitized NMR signal. A signal processing means comprising a plurality of signal processing units for inputting a digital NMR signal output from the A / D conversion means and executing signal processing. There is disclosed a configuration in which each signal processing unit intervenes with the processing means, and the signal processing unit performs different signal processing on the digital NMR signal for each unit and transmits the data to the data processing means in parallel.
JP-A-9-294734

  When an arithmetic unit is provided for each transmission path and NMR signals are processed in parallel, it is necessary to increase the processing efficiency of the signal processing while reducing the cost by reducing the number of transmission paths and the number of arithmetic units as much as possible. One way to increase processing efficiency is to make the data processing load as uniform as possible among multiple computing units in order to avoid a decrease in overall throughput due to an increase in the load on a specific computing unit. Methods for distributing data are generally known.

  However, in the configuration disclosed in (Patent Document 1), data exchange between a plurality of signal processing units is not considered. Further, the data processing means in the subsequent stage is not disclosed regarding parallelization and mutual data exchange. For this reason, when a large amount of data is processed or a large amount of data is generated, the load may be concentrated on a specific signal processing unit. In such a case, the throughput of the entire signal processing system is reduced. It is thought that there is a possibility of decline.

  Accordingly, an object of the present invention is to provide a signal processing system capable of load distribution and increase in processing speed in the case of reconstructing an image by processing an NMR signal received by a multiple coil including a plurality of receiving coils. It is to provide an equipped MRI apparatus.

In order to solve the above problems, the image reconstruction method in the MRI apparatus of the present invention is configured as follows. That is,
A multiple coil having a plurality of receiving coils, and an image reconstructing unit having the same number of transmission paths as the number of the receiving coils and receiving a signal from one receiving coil for each transmission path. An image reconstruction method in a magnetic resonance imaging apparatus comprising:
(a) reconstructing an image of the receiving coil for each transmission path;
(b) a step of combining two images of each transmission path;
(c) synthesizing each of the synthesized images two by two,
(d) repeating the step (c) until the number of the composite images becomes one;
(e) When the number of receiving coils is an odd number, the method further includes a step of combining the last combined image and the image of the remaining transmission path.

This is by reconstructing the image of each receiving coil for each transmission path, combining each image two by two to create a combined image, and then repeating this combined image two by two. Finally, it is a method to synthesize one image. If the number of reception coils is an odd number, the final image is obtained by further combining the last combined image and the remaining transmission path images.
This makes it possible to distribute the load of image reconstruction and increase the processing speed.

Furthermore, in order to solve the above-described problems, the MRI apparatus of the present invention is configured as follows. That is, a multiple coil having a plurality of receiving coils, and an image reconstructing unit having the same number of transmission paths as the number of the receiving coils, and a signal from one receiving coil being input for each transmission path The image reconstructing means has one or more arithmetic units connected in series for each transmission path, and at least a part of the arithmetic units is between the arithmetic units of adjacent transmission paths, In an MRI apparatus connected so as to be able to transmit and receive data in at least one direction, at least one of the one or more computing units in the transmission path is an image from a signal of the receiving coil input to the transmission path. Are reconstructed, and the image of the adjacent computing unit and its own image are synthesized.
Preferably, at least a part of the computing units are interconnected so as to be capable of two-way communication, and send an instruction to start and end the image composition to each other.

This is a configuration of an MRI apparatus that realizes the above image reconstruction method, and by connecting at least a part of a plurality of arithmetic units in a lattice shape to constitute an image reconstruction means, an image of each reception coil is obtained. The reconstruction and the synthesis of these images are distributed to a plurality of computing units.
Thereby, the MRI apparatus of the present invention has the same effect as the image reconstruction method.

  Since the MRI apparatus and the image reconstruction method of the present invention are configured as described above, the transmission paths in the horizontal direction and the vertical direction are determined using the fact that the transmission paths between the arithmetic units are in a lattice shape. It is possible to select a data processing method in consideration, and it is possible to distribute the load of the image reconstruction process to a plurality of arithmetic units. As a result, the data processing speed can be increased, and the overall throughput of the MRI apparatus can be improved.

Hereinafter, embodiments of an MRI apparatus according to the present invention will be described in detail with reference to the drawings.
FIG. 1 is a block diagram showing the overall configuration of an embodiment of an MRI apparatus according to the present invention. This MRI apparatus includes a static magnetic field generation means 1, a gradient magnetic field generation means 2, a transmission system 3, a reception system 4, a sequencer 5, a measurement control device 6, a signal processing arithmetic controller 7, a signal processing arithmetic unit 8, and not shown. It consists of an operation unit and displays a tomographic image of the subject 9 using the NMR phenomenon.

  The static magnetic field generating means 1 has either a permanent magnet, a normal conducting magnet, or a superconducting magnet arranged in a space with a certain circumference around the subject 9, and the body axis direction around the subject 9 or A uniform static magnetic field is generated in a direction perpendicular to the body axis of the subject 9.

  The gradient magnetic field generation system 2 includes a gradient magnetic field coil 10 wound in the three axis directions of X, Y, and Z, and a gradient magnetic field power source 11 that supplies a current to each of these coils. Thus, the gradient magnetic field Gx, Gy, and Gz in the X, Y, and Z axial directions are applied to the subject 9 by supplying current from the gradient magnetic field power supply 11 to each coil. The imaging cross section of the subject 9 is set by applying the gradient magnetic field.

  The transmission system 3 includes a high-frequency oscillator 12, a modulator 13, a high-frequency amplifier 14, and a high-frequency irradiation coil 15, in order to induce an NMR phenomenon in atomic nuclei constituting a predetermined section of the subject 9 In addition, the high-frequency irradiation coil installed close to the subject 9 after the high-frequency signal output from the high-frequency oscillator 12 is modulated by the modulator 13 and amplified by the high-frequency amplifier 14 together with the gradient magnetic field from the gradient magnetic field generating means 2 15, the subject 9 is irradiated with a high frequency magnetic field pulse (hereinafter referred to as “RF pulse”).

  The reception system 4 includes a high-frequency reception coil 16, a reception circuit 17, and an analog / digital (hereinafter referred to as "A / D") converter 18, and is induced by an RF pulse emitted from the high-frequency irradiation coil 15 of the transmission system 3. An NMR signal (echo signal) from the nucleus of the subject 9 is detected by a high-frequency receiving coil 16 disposed close to the subject 9 and input to an A / D converter 18 via a receiving circuit 17 including a preamplifier. And converted into a digital signal. This digital signal is output to the signal processing system 20 as data sampled at the timing controlled by the sequencer 5 (hereinafter referred to as “sampling data”).

  The sequencer 5 operates based on the control of the measurement control device 6, and in order to repeat generation of the RF pulse and the gradient magnetic field based on a predetermined pulse sequence, various commands are given to the gradient magnetic field generation system 2 and the transmission system 3. To control these operations.

  The signal processing system 20 includes a signal processing arithmetic unit 8 and a signal processing arithmetic unit control device 7, and the signal processing arithmetic unit 8 Fourier-transforms the sampling data from the reception system 4 to perform correction calculation, image reconstruction, and the like. The processing is performed, and the processing result is output to the signal processing arithmetic unit controller 7. The signal processing calculator control device 7 displays the image data reconstructed by the signal processing calculator 8 as a tomographic image on the display 19, and stores these image data in a database in the device.

  The measurement control device 6 reads the application pattern of the RF pulse and the gradient magnetic field pulse specified in the designated pulse sequence, operates the sequencer 5, and uses the NMR signal detected by the reception system 4 to use the signal processing system 20 To perform image reconstruction calculation.

(First embodiment)
Hereinafter, a first embodiment of an MRI apparatus according to the present invention will be described. In the present embodiment, the number of receiving coils and the number of transmission paths are an even number.
FIG. 2 is a block diagram illustrating a configuration example of the signal processing arithmetic unit 8 in the MRI apparatus according to the first example of the present embodiment. This signal processing computing unit 8 has a sampling data transmission path corresponding to the number (n) of receiving coils constituting the multiple coil, and an arithmetic unit (hereinafter referred to as `` QD detection and resampling '' for each transmission path. DRF 21), and an arithmetic unit (hereinafter referred to as “RCN”) 22 that performs digital transformation by placing digital data captured from DRF 21 in K space based on position information included in its own header information. Are arranged in series as one arithmetic unit set.

  That is, the above-described arithmetic unit sets are arranged in parallel by a number equal to or greater than the number of receiving coils, sampling data from one receiving coil is input to each transmission path, and parallel processing is performed for each transmission path. As described above, the sampling data is data obtained by digitizing the NMR signal received by the receiving coil connected to the transmission path through the preamplifier and then digitized by the A / D converter.

  At least some of the plurality of arithmetic units are arranged in a lattice pattern and are connected to each other in the vertical direction (that is, the transmission path direction) and the horizontal direction via a data bus. That is, adjacent DRFs 21 and adjacent RCNs 22 are connected to each other, and DRF 21 and RCN 22 are connected in the direction of the transmission path so that data transmission can be performed in one direction or in both directions. For example, a DSP (Digital Signal Processor) can be used as the arithmetic unit. The interconnection between the arithmetic units can exchange data at high speed using a high-speed dedicated data bus (hereinafter simply referred to as “high-speed bus”). For example, an effective data transmission speed of 80 MB / sec or more can be achieved. By connecting at least a part of such a plurality of arithmetic units in a grid using a high-speed bus and efficiently assigning various arithmetic operations to each arithmetic unit, efficient load distribution can be performed. Thereby, the throughput of the entire apparatus can be improved.

FIG. 3 shows an example of the case where the number of receiving coils and the number of transmission paths are an even number, and the flow of data when image reconstruction is performed from NMR signals received by each receiving coil 16 constituting a multiple coil in each case of four. Based on this, a first example of the present embodiment will be described.
The data flow diagram of FIG. 3 shows that for each transmission path, data is transmitted from each receiving coil 16 constituting the multiple coil to the signal processing arithmetic unit 8 via the receiving circuit 17, and the DRF 21 in the signal processing arithmetic unit 8 This shows the processing until the data processed by the RCN 22 is transmitted to the signal processing computing unit controller 7.

  In Step 1, 301 processing is performed for each transmission path. That is, the NMR signal received by each receiving coil 16 is amplified by a preamplifier in the receiving circuit 11, and after passing through the low-pass filter (LPF) 1, the A / D converter converts the NMR signal as an analog signal. It is converted into digitized sampling data. When the sampling rate of the A / D converter is set to 500 kHz, for example, the cutoff frequency of LPF1 is set to about 250 kHz.

  In Step 2, 302 processing is performed for each transmission circuit. That is, sampling data is subjected to quadrature (Q / D) detection and resampling in the DRF 21 in the signal processing arithmetic unit 8. In this quadrature detection, one system of sampling data, which is a digital signal, is digitally quadrature detected and divided into two systems of signals of 0 ° and 90 ° components orthogonal to each other. In resampling, each component of the sampling data is resampled at a sampling rate corresponding to the frequency band of the NMR signal determined by the number of image matrices and the signal measurement time width set at the time of imaging. For example, since the frequency band when measuring 256 matrix data with a signal measurement time width of 5.12 msec is 50 kHz, the sampling rate of resampling is 100 kHz. Therefore, after each component of the sampling data is passed through a digital low-pass filter (LPF) 2 having a cutoff frequency of about 50 kHz, resampling is performed at 100 kHz for each component.

  In Step 3, the process 303 is performed for each transmission path. That is, in the RCN 22 in the signal processing arithmetic unit 8, complex data having a 0 ° component as a real component and a 90 ° component as an imaginary component are arranged in the K space, and all the K space data necessary for image reconstruction are arranged. At that time, complex data is arranged at a position in the K space corresponding to the phase encoding amount applied when the NMR signal is received. After all the K-space data has been prepared, Fourier transformation is performed to reconstruct the image. In this step, the image reconstruction process for each receiving coil 16 is completed.

From the next step, the process of obtaining the final image by combining the images of the receiving coils. This image composition can be performed, for example, as shown in equation (1).
Composite image = (1 / N) Σ (image for each receiving coil) (1)
Here, N is the number of receiving coils.

Accordingly, in Step 4, the process of image composition 1 (304) is performed. The image composition 1 of the expression (1) is performed for every two adjacent transmission paths. For this purpose, for example, the RCN 22-4 of the transmission path 4 transmits the image 4 data to the RCN 22-3 of the adjacent transmission path 3 via the horizontal high-speed bus between the RCNs shown in FIG. The RCN 22-3 adds the image 3 data of its own transmission path 3 and the transmitted image 4 data based on the equation (1).
Composite image 34 = (1/2) {(image 3) + (image 4)} (2)

The same processing is performed for transmission path 1 and transmission path 2. That is, the RCN 22-2 in the transmission path 2 transmits the image 2 data to the RCN 22-1 in the adjacent transmission path 1 via the horizontal high-speed bus between the RCNs shown in FIG. The RCN 22-1 adds the image 1 data of its transmission path 1 and the transmitted image 2 data based on the equation (1).
Composite image 12 = (1/2) {(image 1) + (image 2)} (3)

Finally, in Step 5, image composition 2 (305) is processed. The two synthesized images obtained in the image synthesis 1 are further subjected to image synthesis 2, and a final image is synthesized. For that purpose, the composite image 34 data of the formula (2) synthesized by RCN22-3 of transmission path 3 is transferred to RCN22-1 of transmission path 1 via the horizontal high-speed bus between RCNs shown in FIG. To transmit. At that time, image data is transmitted via the RCN 22-2 of the transmission path 2. In such a case, the internal switch of the RCN 22 is switched to set the horizontal high-speed bus directly. As a result, data transmission can be performed at high speed. RCN 22-1 adds its own composite image 12 data and transmitted composite image 34 data based on the equation (1).
Final composite image = (1/2) {(composite image 12) + (composite image 34)} (4)
The final synthesized image synthesized as described above is transmitted to the signal processing arithmetic unit controller 7.

  Here, it will be briefly described that the data processing of the first embodiment is faster than other possible forms. As a comparison target, it is assumed that images for each transmission path are sequentially added. This case will be described with reference to the configuration of FIG. That is, the RCN 22-3 adds the image 4 reconstructed by the RCN 22-4 and its own image 3 to obtain a composite image 34. Next, in RCN 22-2, the synthesized image 34 synthesized in RCN 22-3 and its own image 2 are added to obtain a synthesized image 234. Finally, in RCN 22-1, the synthesized image 234 synthesized in RCN 22-2 and its own image 1 are added to obtain a final synthesized image. As is clear from this, the number of processing steps when sequentially adding images is larger than the number of processing steps when adding images for every two transmission paths described in the first embodiment. End up. That is, it is clear that the adding method of the first embodiment in which images are added every two transmission paths is efficient.

  In the above description, the example of the lattice arrangement consisting of two stages of DRF and RCN is described as the configuration of the arithmetic unit, but other stages may be used. When the arithmetic unit is composed of only one stage, the DRF 21 in FIG. 2 sequentially performs the processing of the DRF 21 and the RCN 22 described in the present embodiment. That is, after performing from quadrature detection to image reconstruction for each transmission path, the reconstructed image data is transmitted to an adjacent computing unit. The computing unit to which the reconstructed image data has been transmitted can add its own reconstructed image and the transmitted reconstructed image to obtain a composite image. Then, for example, the final composite image is obtained in the leftmost computing unit. Alternatively, in the case of three stages, the second stage performs image reconstruction for each receiving coil as described above, transmits the image data to the third stage computing unit, and the third stage computing unit Image composition can be performed. In the case of four or more stages, the above-described various processes may be further subdivided and distributed to exceptional arithmetic units to perform so-called pipeline processing.

  Next, as a second example of the present embodiment, processing when image reconstruction is performed from NMR signals received by each receiving coil 16 when the number of receiving coils and the number of transmission paths constituting the multiple coil is eight. An example will be described with reference to FIGS.

  FIG. 5 is a block diagram of the internal configuration of the signal processing arithmetic unit 8. In this example, an example in which the number of computing units for each transmission path is one is shown. Sampling data is input to an arithmetic unit (hereinafter referred to as a CH arithmetic unit) for each transmission path, and signal processing is performed. Further, a bus 501 and a bus 502 for performing communication with the measurement control device 6 are connected to each CH computing unit. The bus 501 is a bus capable of performing one-to-one bidirectional communication between the measurement control device 6 and each CH computing unit, and may be a PCI bus, for example. On the other hand, the bus 502 is a bus capable of performing bidirectional communication in parallel between the measurement control device 6 and a plurality of CH computing units and between each CH computing unit, for example, Ethernet (registered trademark). be able to. FIG. 5 shows a case where the bus 502 is Ethernet, and each CH computing unit is connected to the measurement control device 6 via the Ethernet Hub 503 in a star-type network topology and performs bidirectional communication. Is shown. That is, each CH computing unit and Ethernet Hub 503 are connected on a one-to-one basis, and measurement control device 6 and Ethernet Hub 503 are connected by a single line. The Ethernet Hub 503 transmits data from the measurement control device 6 to all the CH calculators, and transmits data from each CH calculator to the measurement control device 6. Each CH computing unit is connected to an adjacent CH computing unit via a high-speed bus 504 that can perform high-speed data transmission only in the lower direction.

FIG. 6 is a schematic block diagram of the internal configuration of each CH computing unit, which includes a CPU, Memory, and Bus Bridge Ship Set. The Bus Bridge Chip Set is connected to the bus 501 and the bus 502, and switches the operation mode of the CH computing unit by controlling a switch in an internal path connected to the high-speed bus.
Sampling data is temporarily stored in the Memory, and the CPU performs each process performed by the RDF and RCN described above on the sampling data on the Memory.
The Bus Bridge Ship Set controls the operation mode of the CH computing unit by controlling the switch. In other words, in Transfer / Receive (T / R) mode, in which the CH arithmetic unit sends and receives data to and from the adjacent CH arithmetic unit, the switch is switched so that the internal path passes through the Bus Bridge Chip Set, and data is simply passed through. In the Through (Thru) mode, the switch is controlled so that the internal path is disconnected from the Bus Bridge Chip Set and directly connected.

7 to 9 are diagrams for explaining an example of image synthesis in the signal processing arithmetic unit configuration of FIG. 5 for each step. Hereinafter, an example of image composition will be described based on these drawings.
FIG. 7 shows the process of image synthesis in the first stage. In the initial state shown in (a), each CH computing unit is in the Transfer / Receive (T / R) mode and performs image reconstruction for each transmission path.

  Next, as shown in the table of (b), in Step 1, the 2CH computing unit transmits image data to the 1CH computing unit via the high-speed bus 504, and the image synthesis processing starts to the 1CH computing unit via the bus 502. The 4CH computing device sends image data to the 3CH computing device via the high-speed bus 504, the 3CH computing device is instructed to start image composition processing via the bus 502, and the 6CH computing device is the high-speed bus. Send image data to the 5CH computing unit via 504, instruct the 5CH computing unit to start image composition processing via the bus 502, and send the image data to the 7CH computing unit via the high-speed bus 504. Instructing the 7CH arithmetic unit to start image composition processing via the bus 502 is performed in parallel. In Step 2, the odd number (that is, 1, 3, 5, 7) CH arithmetic unit performs image data composition processing, and the even number (that is, 2, 4, 6, 8) CH arithmetic unit enters the Through mode. This mode change spontaneously becomes the Through mode after the even-number CH computing unit transfers its own image data in Step 1 described above. In Step 3, the 1CH computing unit notifies the 3CH computing unit, the 3CH computing unit the 5CH computing unit, the 5CH computing unit the 7CH computing unit, and notifies the end of the synthesis process via the bus 502.

  As a result of each of these Step processes, as shown in (c), a composite image obtained by combining the images of the 1CH computing unit and the 2CH computing unit is stored in the 1CH computing unit, and the 3CH computing unit and the 3CH computing unit are combined. A composite image obtained by combining the images of the 4CH calculator is stored, a composite image obtained by combining the images of the 5CH calculator and the 6CH calculator is stored in the 5CH calculator, and the 7CH calculator and the 8CH calculator are combined in the 7CH calculator. A combined image obtained by combining the images is stored. In addition, the 2,4,6CH computing unit is in the Through mode, and the 1,3,5,7CH computing unit remains in the Transfer / Receive mode.

  FIG. 8 shows the image synthesis process in the second stage. The initial state shown in (a) is the same as the end state in the first stage shown in FIG. 7 (c). That is, even-numbered CH arithmetic units are set to the Through mode, and odd-numbered CH arithmetic units are set to the Transfer / Receive mode.

  Next, as shown in the table of (b), in Step 1, the 3CH computing unit transmits image data to the 1CH computing unit via the high-speed bus 504, and the image synthesis processing starts to the 1CH computing unit via the bus 502. The instruction is performed in parallel with the 7CH computing unit transmitting image data to the 5CH computing unit via the high-speed bus 504 and instructing the 5CH computing unit to start the image synthesis process via the bus 502. In Step 2, the 1,5CH computing unit performs image data composition processing, and the 3,7CH computing unit enters the Through mode. This mode change spontaneously becomes the Through mode after the 3,7CH computing unit transfers its own image data in Step 1 described above. In Step 3, the 1CH computing unit notifies the 5CH computing unit of the completion of the synthesis process via the bus 502.

  As a result of each of these Step processes, as shown in (c), a composite image obtained by synthesizing each image of the 1 to 4 CH computing unit is stored in the 1 CH computing unit, and the 5 to 8 CH computing unit is stored in the 5 CH computing unit. A combined image obtained by combining the images is stored. In addition, the 3,7CH arithmetic unit is newly in the Through mode, and only the 1,5CH arithmetic unit remains in the Transfer / Receive mode.

  FIG. 9 shows the process of image synthesis in the third stage. The initial state shown in (a) is the same as the end state of the second stage shown in FIG. 8 (c). That is, the 2-4CH computing unit and the 6-8CH computing unit are set to the Through mode, and the 1,5CH computing unit is set to the Transfer / Receive mode.

Next, as shown in the table of (b), in Step 1, the 5CH computing unit transmits image data to the 1CH computing unit via the high-speed bus 504, and the image synthesis processing starts to the 1CH computing unit via the bus 502. Instruction is given. In Step 2, the 1CH computing unit performs image data composition processing, and the 5CH computing unit enters the Through mode. This mode change spontaneously becomes the Through mode after the 5CH computing unit transfers its own image data in Step 1 described above.
As a result, as shown in (c), the composite image of all the CH calculators is stored in the 1CH calculator. In addition, the 5CH computing unit is newly in the Through mode, and only the 1CH computing unit remains in the Transfer / Receive mode.

  Although the case where the number of transmission paths is four or eight has been described above, an even number other than these may be used. To expand this to the general 2N (N is a natural number), 2M-1 (M is a natural number, M ≦ N) RCN22- (2M-1) th image (2M-1) and image (2M-1) 2M) to obtain a composite image (2M-1,2M). Thereafter, the synthesized images are further synthesized two by two, and the process is repeated until the number of synthesized images becomes one. This last composite image is the final composite image. At this time, the process of synthesizing two synthesized images may be performed in any one of RCNs having synthesized image data.

  As described above, according to the first embodiment of the present invention, when image synthesis is performed using NMR signals received by the respective reception coils constituting the multiple coil, processing is efficiently performed in a short time. Can do. As a result, the number of images that can be reconstructed per unit time can be increased, and a large amount of data generated during real-time imaging can be processed at high speed.

(Second Embodiment)
Next, a second embodiment of the signal processing arithmetic unit 8 will be described. In the present embodiment, the number of receiving coils and the number of transmission paths are odd numbers. Fig. 4 shows the data flow when image reconstruction is performed from NMR signals received by each receiving coil constituting a multiple coil in three cases as an example when the number of receiving coils and the number of transmission paths is odd. Yes, this embodiment will be described based on this. However, the present invention is not limited to the case where there are three transmission paths, but can be applied to other numbers.

The data flow diagram of FIG. 4 shows the data processed by the DRF 21 and RCN 22 in the signal processing unit 8 from the receiving coils 16 constituting the multiple coils via the receiving circuit 17 for each transmission path, as in FIG. Until the signal processing calculator control device 7 is transmitted. The transmission path 4 in FIG. 3 corresponds to the transmission path 3 in FIG. 4, the transmission path 3 in FIG. 3 corresponds to the transmission path 2 in FIG. 4, and the transmission path 1 in FIG. 3 corresponds to the transmission path 1 in FIG. . That is, FIG. 4 shows a case where there is no transmission path corresponding to the transmission path 2 of FIG.
3 is only the transmission path 1, and the processing of the transmission path 2 and the transmission path 3 is the same as that in FIG. Therefore, only the processing of the transmission path 1 will be described and the other description will be omitted.

  In the transmission path 1, since the processing up to the Fourier transform is the same as that in the first embodiment, the description of the processing up to the Fourier transformation is omitted. The image 1 of the transmission path 1 is reconstructed by the Fourier transform. Thereafter, in the present embodiment, the process of image composition 1 (304) in Step 4 is not performed, and only the process of image composition 2 (305) in Step 5 is performed. Accordingly, the composite image 23 synthesized from the image 3 of the transmission path 3 transmitted from the RCN 22-3 and the image 2 of the transmission path 2 of itself is transmitted to the RCN 22-1, and this composite image 23 is transmitted to the RCN 22-1. And the image 1 of its own transmission path 1 are added to obtain a final composite image.

That is, in the present embodiment, a composite image is obtained for each of two adjacent transmission paths, and the final composite image is obtained by adding the own image and another composite image on the remaining one transmission path. Generally speaking, if the number of transmission paths is 2N + 1 (N is a natural number), the image (2M) and the image (2M + 1) are 2M (M is a natural number and M ≦ N). The combined image (2M, 2M + 1) is obtained by combining. Thereafter, the synthesized images are further synthesized two by two, and the process is repeated until the number of synthesized images becomes one. At this time, the process of synthesizing two synthesized images may be performed in any one of RCNs having synthesized image data. Finally, the last synthesized image is transmitted to the RCN of the first transmission path, and the final synthesized image is synthesized by adding the own image (1) and the last synthesized image in the RCN.
In the above description, an example of a lattice arrangement having two stages of DRF and RCN has been described as the configuration of the arithmetic unit. However, the number of other stages may be the same as in the first embodiment described above.

  As described above, according to the second embodiment of the present invention, even when the number of transmission paths is an odd number, the same effect as in the first embodiment is obtained.

The block diagram which shows the whole structure of the MRI apparatus which concerns on this invention. FIG. 3 is a block diagram showing a configuration of a reconstruction calculator 8 according to the present invention. The figure which shows the flow of data in case the number of transmission paths is four. The figure which shows the flow of data in case the number of transmission paths is three. The block diagram which shows another structural example of a signal processing apparatus. FIG. 6 is a block diagram of an internal configuration of a computing unit in the signal processing apparatus of FIG. FIG. 6 is a diagram showing a first stage of a data flow when the number of transmission paths in the signal processing device of FIG. 5 is three. FIG. 6 is a diagram showing a second stage of data flow when the number of transmission paths in the signal processing device of FIG. 5 is three. FIG. 6 is a diagram showing a third stage of data flow when the number of transmission paths in the signal processing device of FIG. 5 is three.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Static magnetic field generation magnet, 2 Static magnetic field gradient generation system, 3 Transmission system, 4 Reception system, 5 Sequencer, 6 Measurement control apparatus, 7 Signal processing calculator control apparatus, 8 Reconstruction calculation machine, 9 Subject, 10 Gradient magnetic field Coil, 11 Gradient magnetic field power supply, 12 High frequency transmitter, 13 Modulator, 14 High frequency amplifier, 15 Irradiation coil, 16 High frequency reception coil, 17 Reception circuit, 18 A / D converter, 19 Display, 20 Signal processing system, 21 DRF , 22 RCN

Claims (2)

An image having a plurality of receiving coils and a plurality of receiving coils, and having the same number of transmission paths as the number of the receiving coils, and receiving a signal from one receiving coil for each transmission path to reconstruct an image. Reconstructing means,
The image reconstruction means has one or more arithmetic units connected in series for each transmission path,
In the magnetic resonance imaging apparatus in which at least a part of the computing unit is connected so as to be able to transmit and receive data in at least one direction between the computing units of adjacent transmission paths.
At least one computing unit among the one or more computing units in the transmission path reconstructs an image from the signal of the receiving coil input to the transmission path, and also includes an image of an adjacent computing unit and its own image. And a magnetic resonance imaging apparatus. The magnetic resonance imaging apparatus according to claim 1 .
At least a part of the arithmetic units are interconnected so as to be capable of two-way communication, and issue a start instruction and an end notification of the image composition to each other.

Images