This repository contains the code for the ICCV 2017 paper "Revisiting Cross-channel Information Transfer for Chromatic Aberration Correction" authored by Tiancheng Sun, Yifan Peng, and Wolfgang Heidrich. Project page
We use Matlab 2016 to implement the algorithm. In the code, we use the open source package mmx and a slightly modified version of im2col. You can find these packages in the folder util/. Please run the code build.m inside to initialize these packages.
We provide an example image taken from diffractive lens. You can use the code test.m to test the algorithm. The core concept is to transfer the information from green channel (which is less burry) to other channels.
We also provide a blind deconvolution version of the green channel (deconvolved using Normalized Sparsity Measure). You can toggle the variable use_blind in test.m to deblur the images.
This code and data is released under the Creative Commons Attribution-NonCommercial 4.0 International license (CC BY-NC.) In a nutshell:
- The license is only for non-commercial use (commercial licenses can be obtained from KAUST, please contact the authors for details)
- The material is provided as-is, with no warranties whatsoever.
- If you publish any code, data, or scientific work based on this, cite our work.
This work was supported by KAUST baseline funding, as well as a UBC 4YF Doctoral Fellowship. The authors thank Tao Yue, Qiang Fu, and Felix Heide for the help on synthetic results.
If you find any bugs or have comments/questions, please contact
Tiancheng Sun [kevin.kingo0627@gmail.com]
Yifan (Evan) Peng [evanpeng@cs.ubc.ca]
Wolfgang Heidrich [wolfgang.heidrich@kaust.edu.sa]
Aug.2017