enricivi / super_resolution

Super-resolution is a technique that constructs an high-resolution image from several observed low-resolution images.

super-resolution

Super-resolution (SR) is a technique that constructs an high-resolution image from several observed low-resolution images. The basic idea behind SR is to combine the non-redundant information contained in multiple low-resolution frames.

Methods implemented on matlab with emmeenne93:

1. Authors: M. Elad, Y. Hel-Or
   Paper: A Fast Super-Resolution Reconstruction Algorithm for Pure Translation Motion and Common Space-Invariant Blur
   
2. Authors: S. Farsiu, M. Dirk Robinson, M. Elad, P. Milanfar
   Paper: Fast and Robust Multiframe Super Resolution

The two papers can be found clicking on Elad and Hel-Or or Milanfar et al.

Run SR.m inside the folder to test the script on "text-test" images.

Note that on Milanfar implementation parameters have to be chosen in order to obtain the best result.