Skip to content

Latest commit

 

History

History
27 lines (24 loc) · 1.63 KB

README.md

File metadata and controls

27 lines (24 loc) · 1.63 KB
Raw

VariationalFrameworkInpainting

Implementation of a variational framework for non-local inpainting in Matlab, based on the paper from Vadim Fedorov, Gabriele Facciolo, Pablo Arias.

This MATLAB implementation of the variational framework for non-local inpainting is made in the context of the final project of the "Traitement d'Images Avancé" (Advance Image Processing) class for Master 2 students, at the University of Bordeaux. It is written by Yiye Jiang, Anais Gastineau and Candice Bentéjac.

To-do list

  • Consider all types of images (grayscale and color images)
  • Write the patch error function E for the patch non-local Poisson
  • Implement the entire patch match algorithm to compute the nearest-neighbor fields
  • Parallelize the patch match algorithm
  • Minimize the energies
    • Correspondance update step
    • Image update step
  • Implement the confidence masks
  • Write the report (easy peasy 😁) -- https://www.overleaf.com/12314065qkxqstqrchft
  • Make it work.

To do (if enough time left)

  • Write the patch error function E for the patch non-local means
  • Write the patch error function E for the patch non-local medians
  • Add the new patch error functions to the patch match algorithm and make them a parameter
  • Add the possibility for the user to choose which patch error function to use
  • Decouple the image and correspondence update steps
  • Implement the multiscale scheme
    • Construction of a masked Gaussian pyramid
    • Joint image and NNF upscaling
  • Add the ability for the user to select the mask by themselves (brush system on the input image)