Spectral image processing method, spectral image processing program, and spectral imaging system
A spectral image processing system and method of performing robust unmixing on measurement noise. Based on an observed spectral image acquired from a specimen and emission spectral data of each of plural materials contained in the specimen, a contribution of each of the plural materials to the observed spectral image is unmixed by a process, including an evaluating step of evaluating reliability of each component of the observed spectral image based on a predicted spectral image of the observed spectral image, and a reflecting step of reflecting a result of the evaluation in a content of the unmixing.
1. A spectral image processing method of, based on an observed spectral image acquired from a specimen and emission spectral data of each of a plurality of materials contained in the specimen, unmixing a contribution of each of said plurality of materials to said observed spectral image, comprising:
an evaluating step of evaluating reliability of each component of said observed spectral image based on a predicted spectral image generated from said observed spectral image by processing image data of said observed spectral image;
a reflecting step of reflecting a result of said evaluation in a content of said unmixing; and
a displaying step of displaying a result of said unmixing.
2. The spectral image processing method according to claim 1 , wherein
said predicted spectral image is a spectral image obtained by smoothing said observed spectral image in a spatial direction.
3. The spectral image processing method according to claim 1 , wherein
in said evaluating step, reliability of said observed spectral image is evaluated with respect to each wavelength component.
4. The spectral image processing method according to claim 1 , wherein
in said evaluating step, reliability of said observed spectral image is evaluated with respect to each wavelength component and each spatial component.
5. The spectral image processing method according to claim 1 , wherein
in said reflecting step, a component, whose said reliability is evaluated as low, of said observed spectral image is excluded from a computation object of said unmixing.
6. The spectral image processing method according to claim 1 , wherein
said unmixing is performed by a weighted least squares method of estimating said contribution after weighting an error of each component of said observed spectral image, and
in said reflecting step, a content of said weighting is set according to the result of said evaluation.
7. The spectral image processing method according to claim 1 , wherein
the evaluating includes comparing each component of said observed spectral image with a corresponding component of said predicted spectral image.
8. The spectral image processing method according to claim 7 , wherein
the components include brightness values.
9. A non-transitory computer-readable storage medium storing a spectral image processing program which causes a computer to execute a spectral image processing method in which a contribution of each of a plurality of materials of a specimen to an observed spectral image acquired from the specimen are unmixed, the spectral image processing method comprising:
an evaluating step of evaluating reliability of each component of said observed spectral image based on a predicted spectral image generated from said observed spectral image by processing image data of said observed spectral image; and
a reflecting step of reflecting a result of said evaluation in a content of said unmixing.
10. The non-transitory computer-readable storage medium according to claim 9 , wherein
the evaluating includes comparing each component of said observed spectral image with a corresponding component of said predicted spectral image.
11. The non-transitory computer-readable storage medium according to claim 10 , wherein
the components include brightness values.
12. A spectral imaging system, comprising:
a spectral imaging unit which acquires an observed spectral image from a specimen; and
a spectral image processing unit which imports said observed spectral image from said spectral imaging unit and performs a spectral image processing method in which contributions of each of a plurality of materials of the specimen to said observed spectral image are unmixed, the spectral image processing method comprising:
an evaluating step of evaluating reliability of each component of said observed spectral image based on a predicted spectral image generated from said observed spectral image by processing image data of said observed spectral image; and
a reflecting step of reflecting a result of said evaluation in a content of said unmixing.
13. The spectral imaging system according to claim 12 , wherein
the evaluating includes comparing each component of said observed spectral image with a corresponding component of said predicted spectral image.
14. The spectral imaging system according to claim 13 , wherein
the components include brightness values.