/Multispectral_Deblurring

Multispectral Deblurring Leveraging Chromatic Aberration for RGB Near-Infrared

Primary LanguageMATLABMIT LicenseMIT

Multispectral Image Deblurring

Majed El Helou, Zahra Sadeghipoor, Sabine Süsstrunk

Contact author: Majed El Helou

Dependencies

The code is tested on MATLAB R2016b and R2018b.

Code

This repository contains the needed MATLAB code to run the functions of "Correlation-Based Deblurring Leveraging Multispectral Chromatic Aberration in Color and Near-Infrared Joint Acquisition". Simply run DEMO_multispectral_deblur.m, it saves intermediate results in the folder ResultsRGBY. To obtain the correlation maps used in the paper's correlation study, use corr2_across_patches.m.

The code allows the deblurring of a spectral channel (red, green, blue, or NIR), given one or more other channels. This takes advantage of the chromatic aberration naturally present in captured images. The code relies on optimizations solved in the Fourier domain (see the second citation below).

The folder Guided_Deblurring contains the demo and code of "Multiscale Guided Deblurring: Chromatic Aberration Correction in Color and Near-Infrared Imaging". Simply run demo.m.

Data

To test with different RGB-NIR image pairs, this dataset can be downloaded. Replace inside Guided_Deblurring the images rgb.tiff and nir.tiff.

Citations

Correlation-Based Deblurring Leveraging Multispectral Chromatic Aberration in Color and Near-Infrared Joint Acquisition:

@inproceedings{el2017correlation,
  title={Correlation-based deblurring leveraging multispectral chromatic aberration in color and near-infrared joint acquisition},
  author={El Helou, Majed and Sadeghipoor, Zahra and S{\"u}sstrunk, Sabine},
  booktitle={International Conference on Image Processing (ICIP)},
  pages={1402--1406},
  year={2017},
  organization={IEEE}
}

Fourier-Domain Optimization for Image Processing:

@article{el2018fourier,
  title={Fourier-domain optimization for image processing},
  author={El Helou, Majed and D{\"u}mbgen, Frederike and Achanta, Radhakrishna and S{\"u}sstrunk, Sabine},
  journal={arXiv preprint arXiv:1809.04187},
  year={2018}
}

Multiscale Guided Deblurring: Chromatic Aberration Correction in Color and Near-Infrared Imaging:

@inproceedings{sadeghipoor2015multiscale,
  title={Multiscale guided deblurring: Chromatic aberration correction in color and near-infrared imaging},
  author={Sadeghipoor, Zahra and Lu, Yue M and Mendez, Erick and S{\"u}sstrunk, Sabine},
  booktitle={European Signal Processing Conference (EUSIPCO)},
  pages={2336--2340},
  year={2015},
  organization={IEEE}
}

Dataset:

@inproceedings{brown2011multi,
  title={Multi-spectral SIFT for scene category recognition},
  author={Brown, Matthew and S{\"u}sstrunk, Sabine},
  booktitle={Computer Vision and Pattern Recognition (CVPR)},
  pages={177--184},
  year={2011},
  organization={IEEE}
}