https://github.com/theveryhim/classic-denoising
Hands on project deploying typical denoising methods.
https://github.com/theveryhim/classic-denoising
image-denoising image-processing
Last synced: 11 months ago
JSON representation
Hands on project deploying typical denoising methods.
- Host: GitHub
- URL: https://github.com/theveryhim/classic-denoising
- Owner: theveryhim
- License: mit
- Created: 2025-07-27T22:44:15.000Z (11 months ago)
- Default Branch: main
- Last Pushed: 2025-07-27T23:59:41.000Z (11 months ago)
- Last Synced: 2025-07-28T01:22:17.399Z (11 months ago)
- Topics: image-denoising, image-processing
- Language: Jupyter Notebook
- Homepage:
- Size: 14.5 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Classic Denoising
In this repo we investigate classic denoising methods.
Our focus is on `Non-Local Means (NLM), BM3D, Total Variation (TV), Weighted Nuclear Norm Minimization (WNNM)`
Find more detailed analysis in notebook!
## Adding Noise
additive Gaussian noise with an average of zero and a deviation from a standard equal to 0.02 of the maximum brightness available in a set of 130 images
## Denoising
```
WNNM PSNR: 16.21 dB
NLM PSNR: 37.32 dB
BM3D PSNR: 33.40 dB
TV PSNR: 33.47 dB
```
## Second order TV
```
Average second_order_tv_psnr: 33.52 dB
```
## Denoising clean image
here we use the *Prob#1.png* image without adding noise as the input of the denoising methods:
```
WNNM PSNR: 30.66 dB
NLM PSNR: 8.06 dB
BM3D PSNR: 111.51 dB
TV PSNR: 8.06 dB
```