neuro_dot.Reconstruction#
Module Contents#
Functions#
|
RECONSTRUCT_IMG Performs image reconstruction by wavelength using the inverted A-matrix. |
|
SMOOTH_AMAT Performs Gaussian smoothing on a sensitivity matrix. |
|
SPECTROSCOPY_IMG Completes the Beer-Lambert law from a reconstructed image. |
|
TIKHONOV_INVERT_AMAT Inverts a sensitivity "A" matrix. |
- neuro_dot.Reconstruction.reconstruct_img(lmdata, iA)#
RECONSTRUCT_IMG Performs image reconstruction by wavelength using the inverted A-matrix.
img = RECONSTRUCT_IMG(data, iA) takes the inverted VOX x MEAS sensitivity matrix “iA” and right-multiplies it by the logmeaned MEAS x TIME light-level matrix “lmdata” to reconstruct an image in voxel space.
The image is output in a VOX x TIME matrix “img”.
See Also: TIKHONOV_INVERT_AMAT, SMOOTH_AMAT, SPECTROSCOPY_IMG, FINDGOODMEAS.
- neuro_dot.Reconstruction.smooth_Amat(iA_in, dim, gsigma)#
SMOOTH_AMAT Performs Gaussian smoothing on a sensitivity matrix.
iA_out = SMOOTH_AMAT(iA_in, dim, gsigma) takes the inverted VOX x MEAS sensitivity matrix “iA_in” and performs Gaussian smoothing in the 3D voxel space on each of the concatenated measurement matrices within, returning it as “iA_out”. The user specifies the Gaussian filter half-width “gsigma”.
See Also: TIKHONOV_INVERT_AMAT, RECONSTRUCT_IMG, FINDGOODMEAS.
- neuro_dot.Reconstruction.spectroscopy_img(cortex_mu_a, E)#
SPECTROSCOPY_IMG Completes the Beer-Lambert law from a reconstructed image.
img_out = SPECTROSCOPY_IMG(img_in, E) takes the reconstructed VOX x TIME x WL mua image “img_in” and multiplies it by the inverse of the extinction coefficient matrix “E” to create an output VOX x TIME x HB matrix “img_out”, where HB 1 and 2 are the voxel-space time series images for HbO and HbR, respectively.
See Also: RECONSTRUCT_IMG, AFFINE3D_IMG.
- neuro_dot.Reconstruction.Tikhonov_invert_Amat(A, lambda1, lambda2)#
TIKHONOV_INVERT_AMAT Inverts a sensitivity “A” matrix.
iA = TIKHONOV_INVERT_AMAT(A, lambda1, lambda2) allows the user to specify the values of the “lambda1” and “lambda2” parameters in the inversion calculation. lambda2 for spatially-variant regularization, is optional.
See Also: SMOOTH_AMAT, RECONSTRUCT_IMG, FINDGOODMEAS.