Signal vector derivation device, method, program, and recording medium
The signal vector deriving device enhances magnetic field measurement accuracy by using multiple sensors and the MUSIC method to derive the direction and position of signal sources with improved precision.
Patent Information
- Application Number
- JP2021074991
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-04-27
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2041-04-27
AI Technical Summary
Existing methods for estimating the position of a signal source from magnetic field measurements lack accuracy.
A signal vector deriving device that utilizes a plurality of sensors to measure components of three orthogonal axes, deriving the direction of a vector based on measurement results, and employing a spectrum deriving unit to determine the direction of the vector using coefficients and the MUSIC method to enhance accuracy.
Improves the measurement accuracy of magnetic fields by precisely determining the direction and position of signal sources, even when the vector is not parallel to the X, Y, or Z axes.
Smart Images

Figure 0007584350000016 
Figure 0007584350000017 
Figure 0007584350000018
Abstract
Description
[Technical field]
[0001] The present invention relates to the measurement of signals such as magnetic fields. [Background technology]
[0002] Conventionally, methods for estimating the position of a signal source from the results of measuring a magnetic field (e.g., subspace methods (MUSIC method, SF, WSF, etc.)) have been known (see Patent Document 1, Non-Patent Document 1, and Non-Patent Document 2). It is also known to measure the magnetic field related to a living body (see Patent Document 2, Patent Document 3, and Patent Document 4). [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Special Publication No. 2009-534103 [Patent Document 2] JP 2009-172088 A [Patent Document 3] JP 2007-20594 A [Patent Document 4] JP 2002-28143 A [Non-Patent Document 1] Ishii H., Niki N., Nakaya Y., Nishitani H., Kang YM (2000) “Algorithm for Magnetocardiography Analysis Based on Multi Channel SQUID and Thoracic MR Images” In: Aine CJ, Stroink G., Wood CC, Okada Y., SwithenbyS.J. (eds) Biomag 96. Springer, New York, NY. pp. 261-262. [Non-Patent Document 2] H. Kudo; T. Maemura; T. Saito “Multiple Signal Source Localization from Spatio-Temporal Magnetocardiogram” Proceedings of 1994 IEEE NuclearScience Symposium - NSS'94 pp. 1832-1836 Summary of the Invention [Problem to be solved by the invention]
[0004] An object of the present invention is to improve the measurement accuracy of signals such as magnetic fields. [Means for solving the problem]
[0005] The signal vector derivation device of the present invention receives a signal represented by a vector having a predetermined direction, and derives the direction of the vector from the measurement results of a plurality of sensors that measure components of three mutually perpendicular axes, wherein the measurement results of the sensors are proportional to the sum of the components of the three axes of the vector multiplied by a first coefficient, and is configured to include a spectrum derivation unit that derives a spectrum that has a maximum value in a voxel where a signal source that outputs the signal is located, the spectrum being obtained based on the measurement results of the sensors and the sum of values obtained by multiplying the first coefficient by a second coefficient, and a direction derivation unit that derives the direction of the vector based on the second coefficient used to obtain the spectrum.
[0006] The signal vector derivation device configured as described above receives a signal represented by a vector having a predetermined direction, and derives the orientation of the vector based on the measurement results of a plurality of sensors measuring components of three mutually orthogonal axes. The measurement results of the sensors are proportional to the sum of the components of the three axes of the vector multiplied by a first coefficient. A spectrum derivation unit derives a spectrum obtained based on the measurement results of the sensors and the sum of values obtained by multiplying the first coefficient by a second coefficient, the spectrum having a maximum value in a voxel where a signal source outputting the signal is present. An orientation derivation unit derives the orientation of the vector based on the second coefficient used to obtain the spectrum.
[0007] In the signal vector derivation device according to the present invention, the vector may be a magnetic dipole moment or an electric dipole moment.
[0008] In the signal vector derivation device according to the present invention, the vector may be an electric dipole moment, and a component of the vector in the same direction as a component of the measurement result may be zero.
[0009] In the signal vector derivation device according to the present invention, the first coefficient may be determined based on a positional relationship between each of the voxels and each of the sensors.
[0010] In the signal vector derivation device according to the present invention, any one or two of the second coefficients may be zero.
[0011] In addition, in the signal vector derivation device according to the present invention, the spectrum may be obtained in accordance with the MUSIC algorithm.
[0012] In the signal vector derivation device according to the present invention, the spectrum derivation section may derive the spectrum based on eigenvectors of a noise subspace obtained from the measurement results of the sensor.
[0013] In the signal vector derivation device according to the present invention, the sum of values obtained by multiplying the first coefficient by the second coefficient may be the transfer function in the MUSIC algorithm.
[0014] In the signal vector derivation device according to the present invention, two or more maximum values may exist.
[0015] The signal vector derivation device according to the present invention may further include a position derivation unit that derives the position of a voxel in which the signal source exists based on the spectrum.
[0016] In the signal vector derivation device according to the present invention, the position derivation section may derive the position of the voxel in which the signal source exists based on the maximum value of each of the spectra in each voxel.
[0017] In addition, in the signal vector derivation device of the present invention, the position derivation unit may determine the center of gravity of voxels that take a value within a predetermined range from the maximum value of the maximum values, while increasing the predetermined range, until the value obtained by adding 1 to the number of times that the center of gravity changes beyond a predetermined amount becomes the number of signal sources, cluster the voxels for which the center of gravity has been determined, the number of clusters being the number of signal sources, and the clustered voxels whose spectrum has the maximum value are determined as the position of the voxel in which the signal source exists.
[0018] In addition, the signal vector derivation device of the present invention may further reduce the size of the voxel based on the position of the voxel in which the signal source exists that has already been derived by the position derivation unit, and derive the position of the voxel in which the signal source exists.
[0019] The present invention is a signal vector derivation method that receives a signal represented by a vector having a predetermined direction, and derives the direction of the vector based on the measurement results of a plurality of sensors that measure components of three mutually perpendicular axes, wherein the measurement results of the sensors are proportional to the sum of values obtained by multiplying each of the three axial components of the vector by a first coefficient, and the signal vector derivation method includes a spectrum derivation step of deriving a spectrum that has a maximum value in a voxel where a signal source that outputs the signal is located, which is obtained based on the measurement results of the sensors and the sum of values obtained by multiplying the first coefficient by a second coefficient, and a direction derivation step of deriving the direction of the vector based on the second coefficient used to obtain the spectrum.
[0020] The present invention is a program for causing a computer to execute a signal vector derivation process that receives a signal represented by a vector having a predetermined direction, and derives the direction of the vector based on the measurement results of a plurality of sensors that measure components of three mutually perpendicular axes, wherein the measurement results of the sensors are proportional to the sum of the components of the three axes of the vector multiplied by a first coefficient, and the signal vector derivation process includes a spectrum derivation step of deriving a spectrum that has a maximum value in a voxel where a signal source that outputs the signal is located, which is obtained based on the measurement results of the sensors and the sum of values obtained by multiplying the first coefficient by a second coefficient, and an orientation derivation step of deriving the direction of the vector based on the second coefficient used to obtain the spectrum.
[0021] The present invention is a computer-readable recording medium having recorded thereon a program for causing a computer to execute a signal vector derivation process that receives a signal represented by a vector having a predetermined direction, receives measurement results from a plurality of sensors that measure components of three mutually perpendicular axes, and derives the direction of the vector, wherein the measurement results from the sensors are proportional to the sum of values obtained by multiplying each of the three axial components of the vector by a first coefficient, and the signal vector derivation process includes a spectrum derivation step of deriving a spectrum that has a maximum value in a voxel where a signal source that outputs the signal is located, which is obtained based on the measurement results from the sensors and the sum of values obtained by multiplying the first coefficient by a second coefficient, and a direction derivation step of deriving the direction of the vector based on the second coefficient used to obtain the spectrum. [Brief description of the drawings]
[0022] [Figure 1] 2 is a perspective view of a voxel V and a magnetic sensor MS according to an embodiment of the present invention. [Diagram 2] 1 is a functional block diagram showing a configuration of a signal vector derivation device 1 according to an embodiment of the present invention. [Diagram 3] 13 is an example of a graph of a maximum value P. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0023] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
[0024] Fig. 1 is a perspective view of a voxel V and a magnetic sensor MS according to an embodiment of the present invention. Fig. 2 is a functional block diagram showing a configuration of a signal vector derivation device 1 according to an embodiment of the present invention.
[0025] Referring to Fig. 1, signal source S1 and signal source S2 output signals. The signals are represented by vector m having a predetermined direction. Vector m is, for example, a magnetic dipole moment. The number of signal sources is, for example, two, but may be three or more as long as it is less than the number of magnetic sensors MS. However, the signals output by each signal source have different frequencies or phases.
[0026] Furthermore, the positions in space where the signal source S1 and the signal source S2 exist are represented by voxels V (for example, 10×10×10=1000 voxels). The signal source S1 and the signal source S2 are each located in a different voxel V. Each of the 1000 voxels V will be denoted as V1 to V1000.
[0027] A plurality of (e.g., 64 in 8 rows and 8 columns) magnetic sensors MS receive a signal (e.g., magnetic dipole moment) and measure components Bx, By, and Bz of three mutually orthogonal axes X, Y, and Z. Each of the 64 magnetic sensors MS is denoted as MS1 to MS64.
[0028] Here, if vector r is the directional vector from the signal source (magnetic dipole) to the magnetic sensor MS, the magnetic flux density B (a function of vector r) measured by the magnetic sensor MS is expressed by Biot-Savart's law as shown in formula (1). Here, μ0 is the magnetic constant. Also, vector r can be said to represent the positional relationship between each of the voxels V (V1 to V1000) and each of the magnetic sensors MS1 to MS64.
[0029]
number
[0030]
number
[0031] From equation (1), By is expressed as the following equation (3).
[0032]
number
[0033] From equation (1), Bz is expressed as the following equation (4).
[0034]
number
[0035] Note that vx1, vx2, vx3, vy1, vy2, vy3, vz1, vz2, and vz3 (first coefficients) are determined based on the vector r with reference to equations (2) to (4) and equations (2') to (4').
[0036] Referring to FIG. 2, the signal vector derivation device 1 according to the embodiment of the present invention includes a relative position recording unit 11, a first coefficient derivation unit 12, a transfer function derivation unit 13, a noise eigenvector derivation unit 14, a spectrum derivation unit 16, a direction derivation unit 18, and a position derivation unit 19.
[0037] The signal vector derivation device 1 receives the measurement results of a plurality of sensors MS1 to MS64 and derives the direction of a vector m.
[0038] The relative position recording unit 11 records vectors r that are the relative positions between each of the 1000 voxels V and each of the magnetic sensors MS1 to MS64.
[0039] The first coefficient derivation unit 12 reads out the vector r from the relative position recording unit 11, and derives the first coefficients vx1, vx2, vx3, vy1, vy2, vy3, vz1, vz2, and vz3 (see equations (2) to (4) and equations (2') to (4')).
[0040] For example, the first coefficient vx1 is expressed as in the following equation (5).
[0041]
number
[0042] Similarly, the other first coefficients vx2, vx3, vy1, vy2, vy3, vz1, vz2, and vz3 each take on 1000 x 64 possible values.
[0043] Although the first coefficient vx1 may be used as is, it is normalized as shown in the following equation (6) before being used in the subsequent processing.
[0044]
number
[0045] The other first coefficients vx2, vx3, vy1, vy2, vy3, vz1, vz2, and vz3 are normalized in the same manner.
[0046] The first coefficient derivation unit 12 outputs the first coefficient normalized as described above.
[0047] The noise eigenvector derivation unit 14 obtains eigenvectors in the noise subspace from the measurement results Bx, By, and Bz of the magnetic sensor MS in accordance with the MUSIC algorithm.
[0048] First, X(t)x is calculated from the measurement result Bx of the magnetic sensor MS as shown in the following formula (7), where t is the time when the measurement is performed and T means transposition.
[0049]
number
[0050] Using X(t)x, the correlation matrix is calculated as shown in the following equation (8).
[0051]
number
[0052] The eigenvector ey of the noise subspace can be found in a similar manner. First, replace Bx in equation (7) with By, and replace X(t)x in equations (7) and (8) with X(t)y to find the correlation matrix using equation (8). Then, similarly, find the eigenvector corresponding to the small eigenvalue and use it as the eigenvector ey of the noise subspace. The eigenvector ey of the noise subspace is a vector with 64 rows and 1 column. There are 62 eigenvectors ey of the noise subspace corresponding to the small eigenvalues.
[0053] The eigenvector ez of the noise subspace can also be found in a similar manner. First, replace Bx in equation (7) with Bz, replace X(t)x in equations (7) and (8) with X(t)z, and find the correlation matrix using equation (8). Then, similarly, find the eigenvector corresponding to the small eigenvalue, and set it as the eigenvector ez of the noise subspace. The eigenvector ez of the noise subspace is a vector with 64 rows and 1 column. There are 62 eigenvectors ez of the noise subspace, corresponding to the small eigenvalues.
[0054] The transfer function derivation unit 13 derives transfer functions vx, vy, and vz as shown in the following formulas (9), (10), and (11). The sum of values obtained by multiplying the first coefficients vx1, vx2, and vx3 by the second coefficients ax, bx, and cx, respectively, is derived (see formula (9)). The derived result is called transfer function vx. The sum of values obtained by multiplying the first coefficients vy1, vy2, and vy3 by the second coefficients ay, by, and cy, respectively, is derived (see formula (10)). The derived result is called transfer function vy. The sum of values obtained by multiplying the first coefficients vz1, vz2, and vz3 by the second coefficients az, bz, and cz, respectively, is derived (see formula (11)). The derived result is called transfer function vz.
[0055]
number
[0056] However, any one or two of the second coefficients may be 0. For example, ax=1, bx=cx=0 (i.e., vx=vx1) or ax=bx=1, cx=0 (i.e., vx=vx1+vx2) may be used.
[0057] Here, there are 13 types of (ak, bk, ck) (where k=x, y, z): (1,0,0), (0,1,0), (0,0,1), (1,1,0), (1,-1,0), (0,1,1), (0,1,-1), (1,0,1), (-1,0,1), (1,1,1), (-1,1,1), (1,-1,1), and (1,1,-1). The vks corresponding to these are vk1, vk2, ..., vk13, respectively.
[0058] For example, if (ax, bx, cx) is (1, 0, 0), then vx = vx1. If (ax, bx, cx) is (1, 1, 0), then vx = vx4 = vx1 + vx2. If (ax, bx, cx) is (1, 1, -1), then vx = vx13 = vx1 + vx2 - vx3.
[0059] For example, if (ay, by, cy) is (1,0,0), then vy=vy1. If (ay, by, cy) is (1,1,0), then vy=vy4=vy1+vy2. If (ay, by, cy) is (1,1,-1), then vy=vy13=vy1+vy2-vy3.
[0060] For example, if (az, bz, cz) is (1, 0, 0), then vz = vz1. If (az, bz, cz) is (1, 1, 0), then vz = vz4 = vz1 + vz2. If (az, bz, cz) is (1, 1, -1), then vz = vz13 = vz1 + vz2 - vz3.
[0061] The spectrum derivation unit 16 derives a spectrum that has a maximum value in the voxel V where the signal sources S1 and S2 exist. Such a spectrum is obtained in accordance with the MUSIC algorithm. The spectrum has two maximum values, corresponding to the number of signal sources. If the number of signal sources is three or more, the spectrum also has three or more maximum values.
[0062] The spectrum is derived by the spectrum derivation unit 16 based on the measurement results Bx, By, Bz (the eigenvectors ex, ey, ez of the noise subspace obtained from them) of the magnetic sensor MS and the sum of the value obtained by multiplying the first coefficient by the second coefficient (i.e., the transfer functions vx, vy, vz) (Equations (9), (10), (11)). The spectrum derivation unit 16 derives the spectrum based on the transfer functions vx, vy, vz output by the transfer function derivation unit 13 and the eigenvectors ex, ey, ez of the noise subspace output by the noise eigenvector derivation unit 14.
[0063] The spectrum derivation unit 16 derives the spectrum Px1 as follows. (1) There are 62 eigenvectors ex (vectors with 64 rows and 1 column) in the noise subspace. These vectors ex are arranged in 62 columns to make ex a matrix with 64 rows and 62 columns. (2) Transpose the matrix ex and multiply it by the transfer function vx1 (a matrix with 64 rows and 1000 columns). That is, ex T Find vx1, which is a matrix with 62 rows and 1000 columns. (3) Square each element of the matrix obtained in (2). (4) The elements of the matrix obtained in (3) are summed column by column and arranged in one row to obtain a matrix with 1 row and 1000 columns. For example, the 1st row, Qth column + the 2nd row, Qth column + ... + the 62nd row, Qth column of the matrix obtained in (3) becomes the element of the 1st row, Qth column of the 1 row, 1000 column matrix obtained in (4) (where Q is an integer between 1 and 1000). (5) By taking the inverse of each element of the matrix obtained in (4), the spectrum Px1 (a 1 row, 1000 column matrix) is obtained.
[0064] Each column of the spectrum Px1 corresponds to voxels V1 to V1000. The same applies to the other spectra.
[0065] The spectrum derivation unit 16 also derives spectra Px2, Px3, ..., Px13. The spectra Px2, Px3, ..., Px13 can be derived by replacing the transfer function vx1 in the above (2) with vx2, vx3, ..., vx13.
[0066] The spectrum derivation unit 16 derives the spectra Py1, Py2, Py3, ..., Py13. By replacing the eigenvector ex of the noise subspace in the above (1) with ey and replacing the transfer function vx1 in the above (2) with vy1, vy2, vy3, ..., vy13, it is possible to derive the spectra Py1, Py2, Py3, ..., Py13.
[0067] The spectrum derivation unit 16 derives the spectra Pz1, Pz2, Pz3, ..., Pz13. By replacing the eigenvector ex of the noise subspace in the above (1) with ez, and replacing the transfer function vx1 in the above (2) with vz1, vz2, vz3, ..., vz13, it is possible to derive the spectra Pz1, Pz2, Pz3, ..., Pz13.
[0068] The position derivation unit 19 derives the position of the voxel V where the signal sources S1 and S2 exist, based on the spectra Px1, Px2, Px3, ..., Px13, Py1, Py2, Py3, ..., Py13, Pz1, Pz2, Pz3, ..., Pz13.
[0069] The spectrum output by the spectrum derivation unit 16 is expressed by the following equation (12).
[0070]
number
[0071]
number
[0072] Fig. 3 is an example of a graph of the maximum value P. In Fig. 3, the vertical axis indicates the spectrum value, and the horizontal axis indicates the voxels (V1 to V1000).
[0073] 3, it is assumed that the value of the maximum value P (of the spectrum) in the voxel V750 is SP1 (maximum value), and the value of the maximum value P (of the spectrum) in the voxel V250 is SP2 (maximum value). However, it is assumed that SP1 is greater than SP2.
[0074] The position derivation unit 19 determines the center of gravity of a voxel whose value falls within a predetermined range from the maximum value SP1 of the maximum values P (for example, the value of maximum value P is 0.95SP1 or more) while increasing the predetermined range (for example, expanding the predetermined range in increments of 0.05SP1, such that the value of maximum value P is 0.95SP1 or more → 0.90SP1 or more → 0.85SP1 or more → ...) until the value obtained by adding 1 to the number of times the center of gravity changes beyond a predetermined amount becomes the number of signal sources (2).
[0075] The center of gravity of the voxels in the vicinity of the maximum value SP1 among the maximum values P is approximately voxel V750. However, when a predetermined range expands from the maximum value SP1 to include SP2, the center of gravity of the voxel becomes considerably smaller than voxel V750. Then, the center of gravity changes beyond a predetermined amount, and the number of times (1 time) plus 1 becomes the number of signal sources (2), so the determination of the center of gravity of the voxel is terminated.
[0076] Next, the voxels whose centroids have been found are clustered by the number of signal sources (2). For example, unsupervised machine learning Kmeans clustering is performed, and the voxels are labeled by the number of signal sources.
[0077] Finally, among the clustered voxels, the one with the maximum spectrum value is determined to be the location of the voxel where the signal source exists.
[0078] The position derivation unit 19 may further reduce the size of the voxel based on the position of the voxel where the signal source exists that has already been derived in this way, to derive the position of the voxel where the signal source exists. In this way, the position of the voxel where the signal source exists can be calculated with high accuracy and high speed.
[0079] The orientation derivation unit 18 receives Pkj (where k=x, y, z and j=1, 2, 3, ...) corresponding to the signal sources S1 and S2 (i.e., taking a maximum value at P) from the position derivation unit 19. Furthermore, the orientation derivation unit 18 derives the orientation of the vector m based on the second coefficient used to obtain Pkj corresponding to the signal sources S1 and S2.
[0080] For example, it is assumed that the position derivation unit 19 provides the orientation derivation unit 18 with Px13 (Py13 or Pz13) in the 750th row (voxel V750) and Px4 (Px4 or Pz4) in the 250th row (voxel V250) as spectra corresponding to signal sources S1 and S2.
[0081] Then, the second coefficients (ak, bk, ck) (where k=x, y, z) used to obtain Px13 (Py13 or Pz13) as the direction of vector m in signal source S1 of voxel V750 are (1, 1, -1). Therefore, the direction derivation unit 18 derives that the direction of vector m in signal source S1 of voxel V750 is parallel to vector (1, 1, -1). However, vector (1, 1, -1) is a vector with an X component of 1, a Y component of 1, and a Z component of -1.
[0082] Furthermore, the second coefficients (ak, bk, ck) (where k=x, y, z) used to obtain Px4 (Px4 or Pz4) as the direction of vector m in signal source S2 of voxel V250 are (1,1,0). Therefore, the direction derivation unit 18 derives that the direction of vector m in signal source S2 of voxel V250 is parallel to vector (1,1,0). However, vector (1,1,0) is a vector with an X component of 1, a Y component of 1, and a Z component of 0.
[0083] Next, the operation of the embodiment of the present invention will be described.
[0084] The first coefficient derivation unit 12 reads out the vector r from the relative position recording unit 11, and derives the first coefficients vx1, vx2, vx3, vy1, vy2, vy3, vz1, vz2, and vz3 (see equations (2) to (4) and equations (2') to (4')).
[0085] The first coefficient has 1000×64 values (see equation (5)), is normalized (see equation (6)), and is provided to the transfer function derivation unit 13.
[0086] The transfer function derivation unit 13 derives transfer functions vx, vy, and vz based on the first and second coefficients ax, bx, cx, ay, by, cy, az, bz, and cz (see equations (9), (10), and (11)).
[0087] The noise eigenvector derivation unit 14 derives eigenvectors ex, ey, and ez of the noise subspace from the measurement results Bx, By, and Bz of the magnetic sensor MS in accordance with the MUSIC algorithm.
[0088] The spectrum derivation unit 16 derives spectra Px1, Px2, Px3, ..., Px13, Py1, Py2, Py3, ..., Py13, Pz1, Pz2, Pz3, ..., Pz13 based on the transfer functions vx, vy, vz and the eigenvectors ex, ey, ez of the noise subspace (see equation (12)).
[0089] The position derivation unit 19 determines the maximum value P of each spectrum in each voxel (i.e., the maximum value in each column of equation (12)) (see equation (13) and FIG. 3). Based on the maximum value P, the voxels 250 and 750 in which the signal sources S1 and S2 exist are derived.
[0090] The direction derivation unit 18 derives the direction of the vector m based on the second coefficient used to obtain Pkj corresponding to the signal sources S1 and S2.
[0091] Embodiments of the present invention improve the accuracy of measurements of signals such as magnetic fields.
[0092] For example, if the transfer function vk includes only vk1, vk2, and vk3, the direction of vector m can be measured only when the direction of vector m is parallel to the X direction, Y direction, or Z direction. If the direction of vector m is parallel to any other direction, for example, the vector (1,1,0) (i.e., a vector with an X component of 1, a Y component of 1, and a Z component of 0), the direction of vector m cannot be measured.
[0093] However, according to an embodiment of the present invention, there are many types of transfer functions vk, such as vk1, vk2, vk3, ..., vk13, so that the direction of vector m can be measured even if the direction of vector m is not parallel to the X direction, Y direction, and Z direction.
[0094] In the embodiment of the present invention, the vector of the signal is assumed to be a magnetic dipole moment, but the vector of the signal is not limited to the magnetic dipole moment. The vector of the signal may be, for example, an electric dipole moment (vector p).
[0095] The magnetic flux density B (a function of vector r) measured by the magnetic sensor MS is expressed as in equation (14).
[0096]
number
[0097]
number
[0098] From equation (14), By is expressed as in equation (16) below.
[0099]
number
[0100] From equation (14), Bz is expressed as in equation (17) below.
[0101]
number
[0102] The configuration and operation of the signal vector derivation device 1 are similar to those in the case where the signal vector is a magnetic dipole moment (vector m), and therefore a description thereof will be omitted.
[0103] The above embodiment can be realized as follows: A computer equipped with a CPU, a hard disk, and a media (USB memory, CD-ROM, etc.) reader reads a medium on which a program for implementing each of the above-mentioned parts, such as the relative position recording unit 11, the first coefficient derivation unit 12, the transfer function derivation unit 13, the noise eigenvector derivation unit 14, the spectrum derivation unit 16, the orientation derivation unit 18, and the position derivation unit 19, is recorded, and the medium is installed on the hard disk. The above-mentioned functions can also be realized by this method. [Explanation of symbols]
[0104] 1. Signal Vector Derivation Device 11 Relative position recorder 12 First coefficient derivation part 13 Transfer function derivation section 14 Noise eigenvector derivation section 16 Spectral derivation section 18 Orientation derivation part 19 Position derivation part MS Magnetic Sensor V Voxel B magnetic flux density vx1, vx2, vx3, vy1, vy2, vy3, vz1, vz2, vz3 First coefficient ak, bk, ck second coefficient vk1, vk2, …, vk13 transfer functions S1, S2 signal source m vector (magnetic dipole moment) ex, ey, ez Eigenvectors of the noise subspace P maximum value
Claims
1. A signal vector derivation device that receives a signal represented by a vector having a predetermined direction, receives measurement results from a plurality of sensors measuring three mutually orthogonal axis components, and derives a direction of the vector, wherein the measurement results from the sensors are proportional to a sum of the three axis components of the vector multiplied by a first coefficient, a spectrum derivation unit that derives a spectrum obtained based on the measurement result of the sensor and a sum of values obtained by multiplying the first coefficient by a second coefficient, the spectrum having a maximum value in a voxel where a signal source that outputs the signal is present; a direction deriving unit that derives a direction of the vector based on the second coefficient used to obtain the spectrum; A signal vector derivation device comprising:
2. 2. The signal vector derivation device according to claim 1, the vector is a magnetic dipole moment or an electric dipole moment, Signal vector derivation device.
3. 2. The signal vector derivation device according to claim 1, the vector is an electric dipole moment, the component of the vector in the same direction as the component of the measurement result is zero; Signal vector derivation device.
4. 2. The signal vector derivation device according to claim 1, The first coefficient is determined based on a positional relationship between each of the voxels and each of the sensors. Signal vector derivation device.
5. 2. The signal vector derivation device according to claim 1, Any one or two of the second coefficients is 0; Signal vector derivation device.
6. 2. The signal vector derivation device according to claim 1, The spectrum is obtained according to the MUSIC method. Signal vector derivation device.
7. 7. A signal vector derivation device according to claim 6, The spectrum derivation unit derives the spectrum based on an eigenvector of a noise subspace obtained from the measurement result of the sensor. Signal vector derivation device.
8. 7. A signal vector derivation device according to claim 6, The sum of values obtained by multiplying the first coefficient by the second coefficient is the transfer function in the MUSIC method. Signal vector derivation device.
9. 2. The signal vector derivation device according to claim 1, There are two or more maximum values. Signal vector derivation device.
10. 2. The signal vector derivation device according to claim 1, A signal vector derivation device comprising a position derivation unit that derives the position of a voxel in which the signal source exists based on the spectrum.
11. 11. A signal vector derivation device according to claim 10, A signal vector derivation device, wherein the position derivation unit derives the position of the voxel in which the signal source exists based on the maximum value of each of the spectra in each voxel.
12. 12. A signal vector derivation device according to claim 11, The position derivation unit, determining the centroids of voxels that have values within a predetermined range from the maximum value among the maximum values while increasing the predetermined range until the number of times that the centroids change by more than a predetermined amount plus one is equal to the number of signal sources; The voxels for which the centroids have been determined are clustered by the number of the signal sources; Among the clustered voxels, the one in which the spectrum has the maximum value is determined as the position of the voxel in which the signal source exists. Signal vector derivation device.
13. 11. A signal vector derivation device according to claim 10, based on the position of the voxel in which the signal source exists that has already been derived by the position derivation unit, the size of the voxel is further reduced to derive the position of the voxel in which the signal source exists; Signal vector derivation device.
14. A signal vector derivation method for receiving a signal represented by a vector having a predetermined direction, and deriving a direction of the vector from measurement results of a plurality of sensors measuring three mutually orthogonal axis components, the measurement results of the sensors being proportional to a sum of the three axis components of the vector multiplied by a first coefficient, a spectrum derivation step of deriving a spectrum obtained based on the measurement result of the sensor and a sum of values obtained by multiplying the first coefficient by a second coefficient, the spectrum having a maximum value in a voxel where a signal source that outputs the signal is present; a direction deriving step of deriving a direction of the vector based on the second coefficient used to obtain the spectrum; A signal vector derivation method comprising:
15. A program for causing a computer to execute a signal vector derivation process that receives a signal represented by a vector having a predetermined direction, receives measurement results from a plurality of sensors that measure three mutually orthogonal axis components, and derives the direction of the vector, the measurement results from the sensors being proportional to the sum of the three axis components of the vector multiplied by a first coefficient, The signal vector derivation process includes: a spectrum derivation step of deriving a spectrum obtained based on the measurement result of the sensor and a sum of values obtained by multiplying the first coefficient by a second coefficient, the spectrum having a maximum value in a voxel where a signal source that outputs the signal is present; a direction deriving step of deriving a direction of the vector based on the second coefficient used to obtain the spectrum; A program with.
16. A computer-readable recording medium having recorded thereon a program for causing a computer to execute a signal vector derivation process for receiving a signal represented by a vector having a predetermined direction, receiving measurement results of a plurality of sensors measuring three mutually orthogonal axis components, and deriving a direction of the vector, wherein the measurement results of the sensors are proportional to a sum of the three axis components of the vector multiplied by a first coefficient; The signal vector derivation process includes: a spectrum derivation step of deriving a spectrum obtained based on the measurement result of the sensor and a sum of values obtained by multiplying the first coefficient by a second coefficient, the spectrum having a maximum value in a voxel where a signal source that outputs the signal is present; a direction deriving step of deriving a direction of the vector based on the second coefficient used to obtain the spectrum; A recording medium comprising:
Citation Information
Patent Citations
Decoherence two-dimensional MUSIC parameter estimation method of electromagnetic vector sensor array
CN106970348A
Cardiac magnetic field diagnostic apparatus of atrial flutter and atrial fibrillation, and method for identifying electrical re-entry circuit of atrial flutter and atrial fibrillation
JP2002028143A
Organism activity analyzer
JP2003038455A
Apparatus for measuring brain function
JP2007020594A
Brain activity analysis method
JP2009172088A
