Medical information processing apparatus, medical information processing method, and storage medium
According to one embodiment, a medical information processing apparatus includes processing circuitry configured to derive an index value with respect to noise included in data associated with magnetic resonance signals collected by each of a plurality of reception coils, adjust a degree to which noise is removed from the data associated with the magnetic resonance signals based on the derived index value, remove noise from the data associated with the magnetic resonance signals based on the adjusted degree, and perform compositing of the data associated with the magnetic resonance signals from which noise has been removed.
1. A medical information processing apparatus comprising:
processing circuitry configured to
acquire a plurality of images of a test object based on a plurality of magnetic resonance signals collected from the test object by a plurality of reception coils, respectively,
derive a plurality of index values with respect to a plurality of noises included in the plurality of images, respectively,
adjust, based on the plurality of index values, a plurality of degrees to which the plurality of noises are removed from the plurality of images, respectively,
remove, based on the plurality of degrees adjusted, the plurality of noises from the plurality of images, respectively,
composite the plurality of images from which the plurality of noises have been removed, respectively, to generate a composite image, and
output the composite image.
2. A medical information processing apparatus comprising:
processing circuitry configured to
acquire a plurality of images of a test object based on a plurality of magnetic resonance signals collected from the test object by a plurality of reception coils, respectively,
composite the plurality of images to generate a composite image,
estimate a noise strength distribution included in the generated composite image,
generate a uniformization distribution for uniformizing the noise strength distribution,
add a signal represented by the uniformization distribution to the composite image to uniformize the noise strength distribution,
remove a noise from the composite image having the uniformized noise strength distribution, and
output the composite image.
3. A medical information processing method for a computer, comprising:
acquiring a plurality of images of a test object based on a plurality of magnetic resonance signals collected from the test object by a plurality of reception coils, respectively;
deriving a plurality of index values with respect to a plurality of noises included in the plurality of images, respectively;
adjusting, based on the plurality of index values, a plurality of degrees to which the plurality of noises are removed from the plurality of images, respectively;
removing, based on the plurality of degrees adjusted, the plurality of noises from the plurality of images, respectively;
compositing the plurality of images from which the plurality of noises have been removed, respectively, to generate a composite image; and
outputting the composite image.
4. A medical information processing method for a computer, comprising:
acquiring a plurality of images of a test object based on a plurality of magnetic resonance signals collected from the test object by a plurality of reception coils, respectively;
compositing the plurality of images to generate a composite image;
estimating a noise strength distribution included in the composite image;
generating a uniformization distribution for uniformizing the noise strength distribution;
adding a signal represented by the uniformization distribution to the composite image to uniformize the noise strength distribution;
removing a noise from the composite image having the uniformized noise strength distribution; and
outputting the composite image.
5. A computer-readable non-transitory storage medium storing a program causing a computer to:
acquire a plurality of images of a test object based on a plurality of magnetic resonance signals collected from the test object by a plurality of reception coils, respectively;
derive a plurality of index values with respect to a plurality of noises included in the plurality of images, respectively;
adjust, based on the plurality of index values, a plurality of degrees to which the plurality of noises are removed from the plurality of images, respectively;
remove, based on the plurality of degrees adjusted, the plurality of noises from the plurality of images, respectively;
composite the plurality of images from which the plurality of noises has been removed, respectively, to generate a composite image; and
outputting the composite image.
6. A computer-readable non-transitory storage medium storing a program causing a computer to:
acquire a plurality of images of a test object based on a plurality of magnetic resonance signals collected from the test object by a plurality of reception coils, respectively;
composite the plurality of images to generate a composite image;
estimate a noise strength distribution included in the composite image;
generate a uniformization distribution for uniformizing the noise strength distribution;
add a signal represented by the uniformization distribution to the composite image to uniformize the noise strength distribution;
remove a noise from the composite image having the uniformized noise strength distribution; and
output the composite image.
7. The medical information processing apparatus according to claim 1 , wherein the processing circuitry is further configured to
derive a plurality of signal-noise ratios of the plurality of images as the plurality of index values, respectively, and
increase the plurality of degrees as the plurality of signal-noise ratios decrease, respectively, and
decrease the plurality of degrees as the plurality of signal-noise ratios increase, respectively.
8. The medical information processing apparatus according to claim 1 , wherein the processing circuitry is further configured to remove the plurality of noises from the plurality of images, respectively, based on a model trained receive a plurality of input images and output a plurality of output images from which a plurality of noises have been removed, respectively.
9. The medical information processing apparatus according to claim 7 , wherein:
the processing circuitry is further configured to derive the plurality of signal-noise ratios based on a plurality of differences in pixel values between a first set of images of the test object and a second set of images of the test object,
the first set of images are acquired while the plurality of reception coils generate a plurality of high-frequency magnetic fields, and
the second set of images are acquired while the plurality of reception coils do not generate the plurality of high-frequency magnetic fields.
10. The medical information processing apparatus according to claim 8 , wherein the processing circuitry is further configured to adjust the plurality of degrees by adjusting an internal parameter of the model.
11. The medical information processing apparatus according to claim 10 , wherein
the internal parameter is a parameter of an activation function of at least one node included in the model.
12. The medical information processing apparatus according to claim 7 , wherein the processing circuitry is further configured to derive the plurality of signal-noise ratios based on a plurality of differences in pixel values between at least two images of the test object acquired while the plurality of reception coils generate a plurality of high-frequency magnetic fields.
13. The medical information processing apparatus according to claim 12 , wherein the at least two images are close to a center slice.
14. The medical information processing apparatus according to claim 2 , wherein the processing circuitry is further configured to estimate the noise strength distribution based on a comparison between the plurality of images and the composite image.
15. The medical information processing method according to claim 3 , further comprising:
deriving a plurality of signal-noise ratios of the plurality of images as the plurality of index values, respectively, and
increasing the plurality of degrees as the plurality of signal-noise ratios decrease, respectively, and
decreasing the plurality of degrees as the plurality of signal-noise ratios increase, respectively, wherein:
deriving the plurality of signal-noise ratios is performed based on a plurality of differences in pixel values between a first set of images of the test object and a second set of images of the test object,
the first set of images are acquired while the plurality of reception coils generate a plurality of high-frequency magnetic fields, and
the second set of images are acquired while the plurality of reception coils do not generate the plurality of high-frequency magnetic fields.
16. The medical information processing method according to claim 3 , further comprising:
deriving a plurality of signal-noise ratios of the plurality of images as the plurality of index values, respectively, and
increasing the plurality of degrees as the plurality of signal-noise ratios decrease, respectively, and
decreasing the plurality of degrees as the plurality of signal-noise ratios increase, respectively,
wherein deriving the plurality of signal-noise ratios is performed based on a plurality of differences in pixel values between at least two images of the test object acquired while the plurality of reception coils generate a plurality of high-frequency magnetic fields.
17. The medical information processing method according to claim 16 , wherein the at least two images are close to a center slice.
18. The medical information processing method according to claim 4 , wherein estimating the noise strength distribution is performed based on a comparison between the plurality of images and the composite image.
19. The computer-readable non-transitory storage medium according to claim 5 , wherein the program further causes the computer to:
derive a plurality of signal-noise ratios of the plurality of images as the plurality of index values, respectively, and
increase the plurality of degrees as the plurality of signal-noise ratios decrease, respectively, and
decrease the plurality of degrees as the plurality of signal-noise ratios increase, respectively, wherein
the plurality of signal-noise ratios are derived based on a plurality of differences in pixel values between a first set of images of the test object and a second set of images of the test object,
the first set of images are acquired while the plurality of reception coils generate a plurality of high-frequency magnetic fields, and
the second set of images are acquired while the plurality of reception coils do not generate the plurality of high-frequency magnetic fields.
20. The computer-readable non-transitory storage medium according to claim 5 , wherein the program further causes the computer to:
derive a plurality of signal-noise ratios of the plurality of images as the plurality of index values, respectively, and
increase the plurality of degrees as the plurality of signal-noise ratios decrease, respectively, and
decrease the plurality of degrees as the plurality of signal-noise ratios increase, respectively,
wherein the plurality of signal-noise ratios are derived based on a plurality of differences in pixel values between at least two images of the test object acquired while the plurality of reception coils generate a plurality of high-frequency magnetic fields.