Highly Accurate Image Reconstruction for Multimodal Noise Suppression Using Semisupervised Learning on Big Data

Abstract:

Impulse noise corruption in digital images frequently occurs because of errors generated by noisy sensors or communication channels, such as faulty memory locations in devices, malfunctioning pixels within a camera, or bit errors in transmission. Although recently developed big data streaming enhances the viability of video communication, visual distortions in images caused by impulse noise corruption can negatively affect video communication applications. Additionally, sparsity, density, and multimodality in large volumes of noisy images have often been ignored in recent studies, whereas these issues have become important because of the increasing viability of video communication services. To effectively eliminate the visual effects generated by the impulse noise from the corrupted images, this study proposes a novel model that uses a devised cost function involving semisupervised learning based on a large amount of corrupted image data with a few labeled training samples. The proposed model qualitatively and quantitatively outperforms the existing state-of-the-art image reconstruction models in terms of the denoising effect.

Existing System:

Numerous filter-based methods have been proposed based on these IN detectors, for example, adaptive center-weighted median filter [8], progressive switching median filter [9], adaptive weighted median filter (AWMF) [10], noise adaptive fuzzy switching median filter (NAFSMF) [11], modified decisionbased unsymmetric trimmed median filter (MDBUTMF) [12], and morphological mean filter (MMF) [13]. These filter-based methods are reasonably effective in eliminating noise artifacts in slightly corrupted images. However, these methods are ineffective for eliminating noise in highly corrupted images (i.e., noise rate higher than 80%) because of the inaccuracy of order statistics.

Proposed System:

The difference of several representative attributes was examined from global, local, and social contexts for image noise removal. These varying attributes are required for a general image reconstruction model in noise removal. The combination power of these representative attributes was clarified by devising a cost function, which was solved by learning the model parameters through a semisupervised learning technique to improve the quantitative and qualitative performance of noise removal. Extensive experiments evidence the improvement and advantage of the proposed method in addressing the problems of Sparsity, Density, and Multimodality in large-volume image sets. A review of the relevant literature revealed that this paper is the first to specifically address the problem of eliminating noise from a large volume of image data corrupted by IN with multimodal density.

CONCLUSIONS:

This paper presented a new image reconstruction model for the IN removal problem. The imaging characteristics of large volumes of noisy images were analyzed. Three categories of features, namely global, local, and social contexts, and their combinations were proposed in the linear model. The calculation of the parameters of the proposed model by using the semisupervised cost function enables the proposed model to robustly manage the reconstruction of noisy images from a large volume of imaging data suffering from the problems regarding Sparsity, Density, and Multimodality. The qualitative and quantitative assessments demonstrated that the proposed model is capable of generating visually pleasing images from large-scale image sets.

REFERENCES:

[1] X. Song, X. Peng, J. Xu, G. Shi, and F. Wu, “Distributed compressive sensing for cloud-based wireless image transmission,” IEEE Transactions on Multimedia, vol. 19, no. 6, pp. 1351–1364, June 2017.

[2] D. A. Huang, L. W. Kang, Y. C. F. Wang, and C. W. Lin, “Self-learning based image decomposition with applications to single image denoising,” IEEE Transactions on Multimedia, vol. 16, no. 1, pp. 83–93, Jan 2014.

[3] D. R. K. Brownrigg, “The weighted median filter,” Commun. ACM, vol. 27, no. 8, pp. 807–818, Aug. 1984. [Online]. Available: http://doi.acm.org/10.1145/358198.358222

[4] T. Chen and H. R. Wu, “Space variant median filters for the restoration of impulse noise corrupted images,” IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing, vol. 48, no. 8, pp. 784–789, Aug 2001.

[5] S. J. Ko and Y. H. Lee, “Center weighted median filters and their applications to image enhancement,” IEEE Transactions on Circuits and Systems, vol. 38, no. 9, pp. 984–993, Sep 1991.

[6] H. Hwang and R. A. Haddad, “Adaptive median filters: new algorithms and results,” IEEE Transactions on Image Processing, vol. 4, no. 4, pp. 499–502, Apr 1995.

[7] Y. , R. H. Chan, and S. Xu, “A detection statistic for random-valued impulse noise,” IEEE Transactions on Image Processing, vol. 16, no. 4, pp. 1112–1120, April 2007.

[8] T. Chen and H. R. Wu, “Adaptive impulse detection using centerweighted median filters,” IEEE Signal Processing Letters, vol. 8, no. 1, pp. 1–3, Jan 2001.

[9] Z. Wang and D. Zhang, “Progressive switching median filter for the removal of impulse noise from highly corrupted images,” IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing, vol. 46, no. 1, pp. 78–80, Jan 1999.

[10] P. Zhang and F. Li, “A new adaptive weighted mean filter for removing salt-and-pepper noise,” IEEE Signal Processing Letters, vol. 21, no. 10, pp. 1280–1283, Oct 2014.

[11] K. K. V. Toh and N. A. M. Isa, “Noise adaptive fuzzy switching median filter for salt-and-pepper noise reduction,” IEEE Signal Processing Letters, vol. 17, no. 3, pp. 281–284, March 2010.

[12] S. Esakkirajan, T. Veerakumar, A. N. Subramanyam, and C. H. PremChand, “Removal of high density salt and pepper noise through modified decision based unsymmetric trimmed median filter,” IEEE Signal Processing Letters, vol. 18, no. 5, pp. 287–290, May 2011.

[13] P. H. Lin, B. H. Chen, F. C. Cheng, and S. C. Huang, “A morphological mean filter for impulse noise removal,” Journal of Display Technology, vol. 12, no. 4, pp. 344–350, April 2016.

[14] J. Xie, R. S. Feris, S. S. Yu, and M. T. Sun, “Joint super resolution and denoising from a single depth image,” IEEE Transactions on Multimedia, vol. 17, no. 9, pp. 1525–1537, Sept 2015.

[15] P. Rodrguez, R. Rojas, and B. Wohlberg, “Mixed gaussian-impulse noise image restoration via total variation,” in 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), March 2012, pp. 1077–1080.