High-Dimensional Mixture Models For Unsupervised Image Denoising
Code related to the paper https://hal.archives-ouvertes.fr/hal-01544249
-
utils.py contains custom functions imread, imshow, imsave, im2pat, pat2im, and psnr
-
HDMI.py contains the HDMI model
-
syntax python run_HDMI_denoising.py image_path --options
replace --options with any:
-s or -stdv standard deviation of the noise -w or --patch_size patch size -n or --n_iter iterations of EM algorithm -k or --n_groups number of groups in the mixture model --gpu to use GPU if possible results --verbose to print time at each iteration