patch_denoise.space_time.utils.eig_analysis#
- patch_denoise.space_time.utils.eig_analysis(input_data, max_eig_val=10)[source]#
Return the eigen values and vectors of the autocorrelation of the patch.
This method is faster than the svd, but the eigen values are in increasing order.
- Parameters:
input_data (np.ndarray) – A 2D Array
max_eig_val (int, optional) – For faster results, only the
max_eig_val
biggest eigenvalues are computed. default = 10
- Returns:
A (numpy.ndarray) – The centered patch A = X - M
d (numpy.ndarray) – The eigenvalues of A^H A
W (numpy.ndarray) – The eigenvector matrix of A^H A
M (numpy.ndarray) – The mean of the patch along the time axis