Skip to content

codes for TNNLS paper "Learning Convolutional Sparse Coding on Complex Domain for Interferometric Phase Restoration"

Notifications You must be signed in to change notification settings

jiankang1991/ComCSC

Repository files navigation

Learning Convolutional Sparse Coding on Complex Domain for Interferometric Phase Restoration

Jian Kang, Danfeng Hong, Jialin Liu, Gerald Baier, Naoto Yokoya, Begüm Demir


This repo contains the codes for the TNNLS paper. we propose a complex convolutional sparse coding (ComCSC) algorithm and its gradient regularized version (ComCSC-GR) for interferometric phase restoration. Our method outperforms the state-of-the-art methods. The codes are modified from SPORCO for adapting to complex images.

alt text 10, 000 Monte-Carlo simulations for evaluating the compared methods on the expected values and standard deviations of step function approximation. The amplitude is constant and the coherence value is set as 0.3.

Usage

train_cpx_dic.m uses ComCSC and ComCSC-GR to train the complex convolutional dictionaries based on the simulated interferograms train_cpxs.mat.

recon_peaks_cpx_comp.mlx gives a demo using the learned convolutional dictionaries to restore the clean interferogram.

Citation

@article{kang2020comcsc,
  title={{Learning Convolutional Sparse Coding on Complex Domain for Interferometric Phase Restoration}},
  author={Kang, Jian and Hong, Danfeng and Liu, Jialin and Baier, Gerald and Yokoya, Naoto and Demir, Begüm},
  journal={IEEE Transactions on Neural Networks and Learning Systems},
  year={2020},
  note={DOI:10.1109/TNNLS.2020.2979546}
  publisher={IEEE}
}

About

codes for TNNLS paper "Learning Convolutional Sparse Coding on Complex Domain for Interferometric Phase Restoration"

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published