from nlmeans import nlmeans import nibabel as nib import os files = ['dsi_cut_v2.nii.gz'] patch_size = 1 nbhood_size = 5 sigma = 0 for img in files: temp, ext = str.split(os.path.basename(img), '.', 1) filename = os.path.dirname(os.path.realpath(img)) + '/' + temp + '_denoised.nii.gz' # '/data/' + temp + '_denoised_nlmeans_rician.nii.gz' denoised = nlmeans(nib.load(img), std=sigma, nbhood_size=patch_size, search_size=nbhood_size) # './data/' + img)) # Patch the b0 to 32767 like it was before denoising data_denoised = denoised.get_data() # data_denoised[...,0] = 32767 img_denoised = nib.Nifti1Image(data_denoised, denoised.get_affine(), denoised.get_header()) nib.save(img_denoised, filename)
from nlmeans import nlmeans import nibabel as nib import os files = ['DWIS_dsi-scheme_SNR-20.nii.gz'] for img in files: temp, ext = str.split(os.path.basename(img), '.', 1) filename = os.path.dirname(os.path.realpath(img)) + '/data/' + temp + '_denoised_nlmeans_rician.nii.gz' denoised = nlmeans(nib.load('./data/' + img)) # Patch the b0 to 32767 like it was before denoising data_denoised = denoised.get_data() data_denoised[...,0] = 32767 img_denoised = nib.Nifti1Image(data_denoised, denoised.get_affine(), denoised.get_header()) nib.save(img_denoised, filename)