Skip to content

Implementation of Fast Image Super-resolution algorithm

Notifications You must be signed in to change notification settings

TopiLeppanen/fisr

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

fisr

Python implementation of Fast Image Super-resolution algorithm

Implements the algorithm in the paper: Fast Image Super-resolution Based on In-place Example Regression, by Jianchao Yang, Zhe Lin, Scott Cohen (2013)

The algorithm needs an external training data set, which is not provided here. For example The Berkeley Segmentation Dataset (500 images) can be used, or any other set of natural images.

This doesn't completely work yet. It doesn't seem to learn the sharpening filters properly. This implementation doesn't implement the "Selective Patch Processing" as described in the paper. Implementing it would have a chance of just speeding up the upscaling part of the algorithm.

About

Implementation of Fast Image Super-resolution algorithm

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages