https://github.com/adrianklessa/svd-image-compression

SVD image compression
https://github.com/adrianklessa/svd-image-compression

data-compression image-compression singular-value-decomposition svd

Last synced: 9 months ago
JSON representation

SVD image compression

• Host: GitHub
• URL: https://github.com/adrianklessa/svd-image-compression
• Owner: AdrianKlessa
• Created: 2023-10-15T14:57:42.000Z (over 2 years ago)
• Default Branch: master
• Last Pushed: 2024-05-01T15:45:06.000Z (almost 2 years ago)
• Last Synced: 2025-06-23T03:44:55.757Z (9 months ago)
• Topics: data-compression, image-compression, singular-value-decomposition, svd
• Language: Python
• Homepage:
• Size: 10 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# SVD-Image-Compression

A simple example of using Singular Value Decomposition for image compression. The example below is compressed by 76% (from a (710, 946, 3) uint8 numpy array to 9 2D arrays using the float64 datatype). 

(2MB to 472 KB)

(710x946x3x8 bits to 71x95x9x64 bits)

Original:

![alt text](test_images/curiosity.jpg)

Compressed:

![alt text](https://github.com/AdrianKlessa/SVD-Image-Compression/blob/b225001415aba2bbfe36c35aaa2e1841d682468c/Compressed%20by%2076.56%20percent.png)

Compression rate is controlled with the `compression_factor` global variable, determining the percentage of singular values to be used.

More compression would likely be achieved with minimal quality loss by forcing usage of lower-precision datatypes than the default float64.