Code for "CURE: Curvature Regularization Via Weighted Nonlocal BiHarmonic For Missing Data Recovery"
Thanks Prof. Zuoqiang Shi and Dr. Wei Zhu for their code for LDMM and WNLL.
- vlfeat-0.9.21 download
- matlab2016b+
(Ground Turth, Inpainting, Sample Rate:20%)
(Ground Turth, W-CUBE:28.56dB, WNLL:27.78dB)
PSNR
SSIM
In our test, we label 700, 100, 70, 50 and 35 images in MNIST respectively. The labeled images are selected at random in 70,000 images. For each sampling rate, we take 10 different random samples for comparisons.
On going work.....
@article{dong2019cure,
title={CURE: Curvature Regularization For Missing Data Recovery},
author={Dong, Bin and Ju, Haocheng and Lu, Yiping and Shi, Zuoqiang},
journal={arXiv preprint arXiv:1901.09548},
year={2019}
}