Robust Image Recovery via Mask Matrix
This paper studies the problem of recovering an unknown image matrix from noisy observations. Existed works, such as Robust Principal Component Analysis (RPCA), are under the case where the image component and error component are additive, but in real world applications, the components are often non-additive. Especially an image may consist of a foreground object overlaid on a background, where each pixel either belongs to the foreground or the background. To separate image components robustly in such a situation, this paper employs a binary mask matrix which shows the location of each component, and proposes a novel image recovery model, called Masked Robust Principal Component Analysis (MaskRPCA). On one hand, the image component and error component are measured by rank function and sparse function, separately. On another hand, the non-additive between components is characterized by mask matrix. Then we develop an iterative scheme based on alternating direction method of multipliers. Extensive experiments on face images and videos demonstrate the effectiveness of the proposed algorithm.
KeywordsNon-additive signal Low rank Sparse Robust image recovery
The authors would like to thank the anonymous reviewers for their valuable comments and suggestions to improve the quality of the paper. This work was supported in part by the National Nature Science Foundation of China 61672291 and Six talent peaks project in Jiangsu Province SWYY-034.
- 3.Wright, J., Ganesh, A., Rao, S., Ma, Y.: Robust principal component analysis: exact recovery of corrupted low-rank matrices by convex optimization. In: Proceedings of the 22nd International Conference on Neural Information Processing Systems, pp. 2080–2088. MIT Press, Vancouver (2009)Google Scholar
- 13.Yi, X., Park, D., Chen, Y., Caramanis, C.: Fast algorithms for robust PCA via gradient descent. In: International Conference on Neural Information Processing Systems, pp. 4152–4160. MIT Press, Barcelona (2016)Google Scholar
- 16.Lin, Z.C., Chen, M.M., Ma, Y.: The augmented Lagrange multiplier method for exact recovery of corrupted low-rank matrix. Technical Report UILU-ENG-09–2215, UIUC, October 2009Google Scholar
- 18.Nie, F., Yuan, J., Huang, H.: Optimal mean robust principal component analysis. In: 31st International Conference on Machine Learning, pp. 1062–1070. MIT Press, Beijing (2014)Google Scholar
- 19.Netrapalli, P., Niranjan, U.N., Sanghavi, S.: Provable non-convex robust PCA. In: International Conference on Neural Information Processing Systems, pp. 1107–1115. MIT Press, Montreal (2014)Google Scholar
- 20.Zhou, P., Feng, J.: Outlier-robust tensor PCA. In: IEEE Conference on Computer Vision and Pattern Recognition, pp. 3938–3946. IEEE, Honolulu (2017)Google Scholar