dcmri.deconv#

dcmri.deconv(h, g, dt=1.0, order=2, method='TSVD', tol=1e-15)[source]#

Deconvolve two uniformly sampled 1D functions.

If and (h,g) are known in h = g*f, this function estimates f = deconv(h, g).

Parameters:
  • h (numpy array) – Result of the convolution. if g has N elements, than h can be an N-element array, or a N x K - element array where N is the length of g. In this case each column is deconvolved indepdently with g.

  • g (numpy array) – One factor of the convolution (1D array).

  • dt (float, optional) – Time between samples. Defaults to 1.0.

  • order (int, optional) – Integration order of the convolution matrix. Defaults to 2.

  • method (str, optional) – Regularization method. Current options are ‘TSVD’ (Truncated Singular Value Decomposition) or ‘Tikhonov’. Defaults to False.

  • tol (float, optional) – Tolerance for the inversion of the convolution matrix (rgularization parameter). Singular values less than a fraction ‘tol’ of the largest singular value are ignored. Defaults to 1e-15.

Returns:

Estimate of the convolution factor f. This has the same shape as h.

Return type:

numpy.ndarray