Image Grain (Low-Frequency Noise) Removal Using White Noise Filter

This is an implementation of LFNFilter. on Python and Matlab. See also LFNFilter. LFNFilter uses existing white Gaussian noise filters (e.g., BMCNN or
CMDF) to address low-frequency component of spatially correlated noise (image grain). Here is the block diagram of the LFNFilter.

Here is a sample denoising result using BMCNN. Methods designed to remove AWGN fail on removing grain noise (left). Same method can be used in LFN architecture to remove grain (right).

The repository includes:

• Todo: Python package.
• Matlab package for LFNFilter.
• Test images.
• Trained weights for BMCNN.

Getting Started with Python (Todo)

Getting Started with Matlab (Windows only)

• demo_LFN.m Is the fastest way to start.
• demo_LFN_real.m demonstrates the real image grain removal.

Installation (Python)

1. Todo