Color image completion using tensor truncated nuclear norm with l0 total variation

Karima EL Qate, Souad Mohaoui, Abdelilah Hakim, Said Raghay

Abstract


In recent years, the problem of incomplete data has been behind the appearance of several completion methods and algorithms. The truncated nuclear norm has been known as a powerful low-rank approach both for the matrix and the tensor cases. However, the low-rank approaches are unable to characterize some additional information exhibited in data such as the smoothness or feature-preserving properties. In this work, a tensor completion model based on the convex truncated nuclear norm and the nonconvex-sparse total variation is introduced. Notably, we develop an alternating minimization algorithm that combines the accelerating proximal gradient for the convex step and a projection operator for the nonconvex step to solve the optimization problem. Experiments and comparative results show that our algorithm has a significant impact on the completion process.


Full Text:

PDF

References


Y. Hu, D. Zhang, J. Ye, X. Li, and X. He, Fast and accurate matrix completion via truncated nuclear norm regularization, IEEE transactions on pattern analysis and machine intelligence 35 9 (2012), no. 9, 2117-2130. DOI: 10.1109/TPAMI.2012.271

T.-Y. Ji, T.-Z. Huang, X.-L. Zhao, T.-H. Ma, and G. Liu, Tensor completion using total variation and low-rank matrix factorization, Information Sciences 326 (2016), 243-257. DOI: 10.1016/j.ins.2015.07.049

Q. Liu, Z. Lai, Z. Zhou, F. Kuang, and Z. Jin, A truncated nuclear norm regularization method based on weighted residual error for matrix completion, IEEE Transactions on Image Processing 25 (2015), no. 1, 316-330. DOI: 10.1109/TIP.2015.2503238

S. Mohaoui, A. Hakim, and S. Raghay, Tensor completion via bilevel minimization with fixed-point constraint to estimate missing elements in noisy data, Advances in Computational Mathematics 47 (2021), no. 1, 1-27. DOI: 10.1007/s10444-020-09841-8

Y. Nesterov, Gradient methods for minimizing composite functions, Mathematical Programming 140 (2013), no. 1, 125-161. DOI: 10.1007/s10107-012-0629-5

S. Ono, L0 gradient projection, IEEE Transactions on Image Processing 26 (2017), no. 4, 1554-1564. DOI: 10.1109/TIP.2017.2651392

Y. Song, J. Li, X. Chen, D. Zhang, Q. Tang, and K. Yang, An efficient tensor completion method via truncated nuclear norm, Journal of Visual Communication and Image Representation 70 (2020), 102791. DOI: 10.1016/j.jvcir.2020.102791

M. Wang, Q. Wang, and J. Chanussot, Tensor low-rank constraint and l0 total variation for hyperspectral image mixed noise removal, IEEE Journal of Selected Topics in Signal Processing 15 (2021), no. 3, 718-733. DOI: 10.1109/JSTSP.2021.3058503

J. Wright, A. Ganesh, S. Rao, and Y. Ma, Robust principal component analysis: Exact recovery of corrupted low-rank matrices via convex optimization, Coordinated Science Laboratory Report no. UILU-ENG-09-2210, DC-243 (2009).

S. Xue, W. Qiu, F. Liu, and X. Jin, Low-rank tensor completion by truncated nuclear norm regularization, In: 2018 24th International Conference on Pattern Recognition(ICPR), IEEE, 2600-2605, (2018).

D. Zhang, Y. Hu, J. Ye, X. Li, and X. He, Matrix completion by truncated nuclear norm regularization, In: 2012 IEEE Conference on computer vision and pattern recognition, IEEE, 2192-2199, (2012).




DOI: https://doi.org/10.52846/ami.v49i2.1532