Fast image noise estimation (Estimation of Gaussian, signal-dependent, and processed noise in Image and Video Signals)
This is an implementation of IVHC on Python and Matlab. See also IVHC. IVHC is a model to estimate Gaussian, signal-dependent, and processed noise in image and video signals. The estimation is based on the classification of intensity-variances of image patches in order to find homogeneous regions that best represent the noise.
Here is the block diagram of the intensity-variance homogeneity classification (IVHC) noise estimation.
Inputs:
Outputs:
The repository includes:
Python Installation
Install dependencies pip3 install package [numpy, scikit-mage, ...]
Run setup from the libs directory python3 setup.py install optional:
Run demos.py: python3 demos.py
demo_awgn.m is the easiest way to start. It shows an example of estimating AWGN.
demo_pgn.m PGN (signal-dependent) noise estimation.
demo_ppn.m PPN processed noise estimation.
demo_real.m non-synthetic (real) image noise.
demo_compare_awgn.m compare AWGN with other method.
demo_compare_ppn.m compare PPN with other method.