access icon free Deep convolutional neural network for mixed random impulse and Gaussian noise reduction in digital images

© The Institution of Engineering and Technology
Received 23/07/2019, Accepted 06/11/2020, Revised 30/09/2020, Published 04/12/2020

This study utilises a deep convolutional neural network (CNN) implementing regularisation and batch normalisation for the removal of mixed, random, impulse, and Gaussian noise of various levels from digital images. This deep CNN achieves minimal loss of detail and yet yields an optimal estimation of structural metrics when dealing with both known and unknown noise mixtures. Moreover, a comprehensive comparison of denoising filters through the use of different structural metrics is provided to highlight the merits of the proposed approach. Optimal denoising results were obtained by using a 20-layer network with 40 × 40 patches trained on 400 180 × 180 images from the Berkeley segmentation data set (BSD) and tested on the BSD100 data set and an additional 12 images of general interest to the research community. The comparative results provide credence to the merits of the proposed filter and the comprehensive assessment of results highlights the novelty and performance of this CNN-based approach.

Inspec keywords: filtering theory; Gaussian noise; image classification; learning (artificial intelligence); image denoising; image segmentation; convolution; neural nets

Other keywords: digital images; optimal estimation; mixed impulse; convolutional neural network; deep CNN; unknown noise mixtures; known noise mixtures; optimal denoising results; Gaussian noise reduction; different structural metrics; minimal loss; additional 12 images; 20-layer network; batch normalisation; CNN-based approach; mixed random impulse

Subjects: Neural computing techniques; Computer vision and image processing techniques; Knowledge engineering techniques; Filtering methods in signal processing; Optical, image and video signal processing

References

    1. 1)
      • 25. Arnal, J., Sanchez, M.G., Vidal, V.: ‘Parallel filter for mixed Gaussian-impulse noise removal’. IEEE Int. Conf. Signal Processing: Algorithms, Architectures, Arrangements, and Applications, Poznan, Poland, 2013.
    2. 2)
      • 16. Lian, D., Hu, L., Luo, W., et al: ‘Multiview multitask gaze estimation with deep convolutional neural networks’, IEEE Trans. Neural Netw. Learn. Syst., 2018, 30, (10), pp. 3010–3023.
    3. 3)
      • 22. Su, T.J., Li, C.I.: ‘An adaptive filtering method for mixed noise of images’. IEEE Int. Symp. on Computer, Consumer and Control, Taichung, Taiwan, 2012.
    4. 4)
      • 63. He, K., Zhang, X., Ren, S., et al: ‘Delving deep into rectifiers: surpassing human-level performance on imagenet classification’. Proc. IEEE Conf. Computer Vision, Las Condes, Chile, 2015, pp. 1026–1034.
    5. 5)
      • 20. Buades, A., Coll, B., Morel, J.-M.: ‘A non-local algorithm for image denoising’. Proc. IEEE Int. Conf. Computer Vision and Pattern Recognition, San Diego, CA, USA, 2005, vol. 2, pp. 60–65.
    6. 6)
      • 37. Delon, J., Desolneux, A., Guillemot, T.: ‘PARIGI: a patch-based approach to remove impulse-Gaussian noise from images’, Image Process. On Line, 2014, 5, pp. 130–154.
    7. 7)
      • 27. Jayasree, M., Narayanan, N.K.: ‘A novel fuzzy filter for mixed impulse Gaussian noise from color images’. Proc. Int. Conf. Signal, Networks, Computing, and Systems, New Delhi, India, 2016, pp. 53–59.
    8. 8)
      • 8. Hamouda, M., Ettabaa, K.S., Bouhlel, M.S.: ‘Hyperspectral imaging classification based on convolutional neural networks by adaptive sizes of windows and filters’, IET Image Proc., 2019, 13, (2), pp. 392–398.
    9. 9)
      • 34. Jiang, J., Yang, J., Cui, Y., et al: ‘Mixed noise removal by weighted low rank model’, Neurocomputing, 2015, 151, pp. 817–826.
    10. 10)
      • 4. Liu, W., Lin, W.: ‘Additive white Gaussian noise level estimation in SVD domain for images’, IEEE Trans. Image Process., 2013, 22, (3), pp. 872–883.
    11. 11)
      • 60. Zhao, H., Gallo, O., Frosio, I., et al: ‘Loss functions for image restoration with neural networks’, IEEE Trans. Comput. Image., 2017, 3, (1), pp. 47–57.
    12. 12)
      • 41. Shen, Y., Han, B., Braverman, E.: ‘Removal of mixed Gaussian and impulse noise using directional tensor product complex tight framelets’, J. Math. Imag., 2016, 54, (1), pp. 64–77.
    13. 13)
      • 51. Zhang, L., Zhang, L., Mou, X., et al: ‘FSIM: A feature similarity index for image quality assessment’, IEEE Trans. Image Process., 2011, 20, (8), pp. 2378–2386.
    14. 14)
      • 40. Mendiola-Santibañez, J.D., Terol-Villalobos, I.R.: ‘Filtering of mixed Gaussian and impulsive noise using morphological contrast detectors’, IET Image Process., 2014, 8, (3), pp. 131–141.
    15. 15)
      • 38. Aher, R.P., Jodhanle, K.C.: ‘Removal of mixed impulse noise and Gaussian noise using genetic programming’. Proc. IEEE Int. Conf. Signal Processing, Beijing, People's Republic of China, 2012.
    16. 16)
      • 47. Cai, J.F., Chan, R., Nikolova, M.: ‘Fast two-phase image deblurring under impulse noise’, J. Math. Imag. Vis., 2010, 36, pp. 46–53.
    17. 17)
      • 21. Garnett, R., Huegerich, T., Chui, C., et al: ‘A universal noise removal algorithm with an impulse detector’, IEEE Trans. Image Process., 2005, 14, (11), pp. 1747–1754.
    18. 18)
      • 54. Jain, V., Seung, S.: ‘Natural image denoising with convolutional networks’. Proc. Conf. Advances in Neural Information Processing Systems (NIPS), Vancouver, BC, Canada, 2009, vol. 21, pp. 769–776.
    19. 19)
      • 30. Chankhachon, S., Intajag, S.: ‘Resourceful method to remove mixed Gaussian-impulse noise in color images’. 12th IEEE Int. Conf. Computer Science and Software Engineering, Songkhla, Thailand, 2015.
    20. 20)
      • 57. Ioffe, S., Szegedy, C.: ‘Batch normalization: accelerating deep network training by reducing internal covariate shift’. Proc. Conf. Machine Learning (ICML), Lille, France, 2015, pp. 448–456.
    21. 21)
      • 13. Liu, N., Han, J., Liu, T., et al: ‘Learning to predict eye fixations via multiresolution convolutional neural networks’, IEEE Trans. Neural Netw. Learn. Syst., 2018, 29, (2), pp. 392–404.
    22. 22)
      • 36. Eslahi, N., Mahdavinataj, H., Aghagolzadeh, A.: ‘Mixed Gaussian-impulse noise removal from highly corrupted images via adaptive local and nonlocal statistical priors’. 9th IEEE Iranian Conf. Machine Vision and Image Processing, Tehran, Iran, 2015.
    23. 23)
      • 10. Khaw, H.Y., Soon, F.C., Chuah, J.H., et al: ‘High-density impulse noise detection and removal using deep convolutional neural network with particle swarm optimisation’, IET Image Proc., 2019, 13, (2), pp. 365–374.
    24. 24)
      • 58. Levin, A., Nadler, B.: ‘Natural image denoising: optimality and inherent bounds’. Proc. IEEE Conf. Computer Vision and Pattern Recognition, Providence, RI, USA, 2011, pp. 2833–2840.
    25. 25)
      • 11. Chen, Y., Pock, T.: ‘Trainable nonlinear reaction diffusion: a flexible framework for fast and effective image restoration’, IEEE Trans. Pattern Anal. Mach. Intell., 2017, 39, (6), pp. 1256–1272.
    26. 26)
      • 32. Filipović, M., Jukić, A.: ‘Restoration of images corrupted by mixed Gaussian-impulse noise by iterative sost-hard thresholding’, Pattern Recognit., 2011, 44, (8), pp. 1708–1720.
    27. 27)
      • 48. Mafi, M., Martin, H., Adjouadi, M.: ‘High impulse noise intensity removal in MRI images’. IEEE Symp. on Signal Processing in Medicine and Biology Symp., Philadelphia, PA, USA, 2017.
    28. 28)
      • 3. Liu, C., Szeliski, R., Kang, S.B., et al: ‘Automatic estimation and removal of noise from a single image’, IEEE Trans. Pattern Anal. Mach. Intell., 2008, 30, (2), pp. 299–314.
    29. 29)
      • 26. Morillas, S., Gregori, V., Hervás, A.: ‘Fuzzy peer groups for reducing mixed Gaussian-impulse noise from color images’, IEEE Trans. Image Process., 2009, 18, (7), pp. 1452–1466.
    30. 30)
      • 53. Burger, H.C., Schuler, C.J., Harmeling, S.: ‘Image denoising: can plain neural networks compete with BM3D?’. Proc. IEEE Conf. Computer Vision and Pattern Recognition, Providence, RI, USA, 2012, pp. 2392–2399.
    31. 31)
      • 14. Shi, W., Gong, Y., Tao, X., et al: ‘Training DCNN by combining max-margin, max-correlation objectives, and correntropy loss for multilabel image classification’, IEEE Trans. Neural Netw. Learn. Syst., 2018, 29, (7), pp. 2896–2908.
    32. 32)
      • 35. Huang, T., Dong, W., Xie, X., et al: ‘Mixed noise removal via Laplacian scale mixture modeling and nonlocal low-rank approximation’, IEEE Trans. Image Process., 2017, 26, (7), pp. 3171–3186.
    33. 33)
      • 9. Zhang, F., Cai, N., Wu, J., et al: ‘Image denoising method based on a deep convolution neural network’, IET Image Proc., 2018, 12, (4), pp. 485–493.
    34. 34)
      • 28. Sun, Y., Junwei, H., Jun, L.: ‘An information-fusion edge preserving method in image filtering’. 16th IEEE Conf. Wireless Communications Networking and Mobile Computing, Chengdu, People's Republic of China, 2010.
    35. 35)
      • 62. Kingma, D., Ba, J.: ‘Adam: A method for stochastic optimization’. Conf. Learning Representations, San Diego, CA, USA, 2015.
    36. 36)
      • 44. Dabov, K., Foi, A., Katkovnik, V., et al: ‘Image denoising by sparse 3-d transform-domain collaborative filtering’, IEEE Trans. Image Process., 2007, 16, (8), pp. 2080–2095.
    37. 37)
      • 46. Xiong, B., Yin, Z.: ‘A universal denoising framework with a new impulse detector and nonlocal means’, IEEE Trans. Image Process, 2012, 21, (4), pp. 1663–1675.
    38. 38)
      • 7. Shamsolmoali, P., Zareapoor, M., Yang, J.: ‘Convolutional neural network in network (CNNiN): hyperspectral image classification and dimensionality reduction’, IET Image Proc., 2019, 13, (2), pp. 246–253.
    39. 39)
      • 29. Guo, X., Guo, B.: ‘A fuzzy filter for color images corrupted by mixed noise’. IEEE Int. Conf. Identification, Information and Knowledge in the Internet of Things, Beijing, People's Republic of China, 2014.
    40. 40)
      • 19. Bing, L., QuanSheng, L., JiaWei, X., et al: ‘A new method for removing mixed noises’, Sci. Chin. Inf. Sci., 2011, 54, (1), pp. 51–59.
    41. 41)
      • 23. Huang, Y.M., Ng, M.K., Wen, Y.W.: ‘Fast image restoration methods for impulse and Gaussian noises removal’, IEEE Signal Process. Lett., 2009, 16, (6), pp. 457–460.
    42. 42)
      • 59. Simonyan, K., Zisserman, A.: ‘Very deep convolutional networks for large-scale image recognition’. Conf. Learning Representations, San Diego, CA, USA, 2015.
    43. 43)
      • 43. Ji, H., Huang, S., Shen, Z., et al: ‘Robust video restoration by joint sparse and low rank matrix approximation’, SIAM J. Imag. Sci., 2011, 4, (4), pp. 1122–1142.
    44. 44)
      • 55. Ding, L., Zhang, H., Xiao, J., et al: ‘An improved image mixed noise removal algorithm based on superresolution algorithm and CNN’, Neural Comput. Appl., 2019, 31, (Suppl 1), pp. S325–S336.
    45. 45)
      • 45. Dong, W., Zhang, L., Shi, G., et al: ‘Nonlocally centralized sparse representation for image restoration’, IEEE Trans. Image Process., 2013, 22, (4), pp. 1620–1630.
    46. 46)
      • 56. Krizhevsky, A., Sutskever, I., Hinton, G.E.: ‘Image net classification with deep convolutional neural networks’. Proc. Conf. Advances in Neural Information Processing Systems (NIPS), Lake Tahoe, NV, USA, 2012, pp. 1097–1105.
    47. 47)
      • 24. Rodríguez, P., Rojas, R., Wohlberg, B.: ‘Mixed Gaussian-impulse noise image restoration via total variation’. IEEE Int. Conf. Acoustics, Speech, and Signal processing, Kyoto, Japan, 2012.
    48. 48)
      • 50. Healey, G.E., Kondeputy, R.: ‘Radiometric CCD camera calibration and noise estimation’, IEEE Trans. Pattern Anal. Mach. Intell., 1994, 16, (3), pp. 267–276.
    49. 49)
      • 42. Cai, J., Chan, R.H., Nikolova, M.: ‘Two-phase methods for deblurring images corrupted by impulse plus Gaussian noise’, Inverse Probl. Imag., 2008, 2, pp. 187–204.
    50. 50)
      • 1. Bovik, A.C.: ‘Handbook of image and video processing’ (Academic press, Cambridge, MA, USA, 2000, 2nd edn.).
    51. 51)
      • 6. Xu, Y., Zhou, X., Chen, S., et al: ‘Deep learning for multiple object tracking: a survey’, IET Comput. Vis., 2019, 13, (4), pp. 355–368.
    52. 52)
      • 31. Xiao, Y., Zeng, T., Yu, J., et al: ‘Restoration of images corrupted by mixed Gaussian-impulse noise via l1–l0 minimization’, Pattern Recognit., 2011, 44, pp. 1708–1720.
    53. 53)
      • 49. Dong, Y., Chan, R.H., Xu, S.: ‘A detection statistic for random-valued impulse noise’, IEEE Trans. Image Process., 2007, 16, (4), pp. 1112–1120.
    54. 54)
      • 39. Lamichhane, B.P.: ‘Finite element techniques for removing the mixture of Gaussian and impulsive noise’, IEEE Trans. Signal Process., 2009, 57, (7), pp. 2538–2547.
    55. 55)
      • 18. Mafi, M., Tabarestani, S., Cabrerizo, M., et al: ‘Denoising of ultrasound images affected by combined speckle and Gaussian noise’, IET Image Proc., 2016, 12, (12), pp. 2346–2351.
    56. 56)
      • 17. Mafi, M., Rajaei, H., Cabrerizo, M., et al: ‘A robust edge detection approach in the presence of high impulse noise intensity through switching adaptive median and fixed weighted mean filtering’, IEEE Trans. Image Process., 2018, 27, (11), pp. 5475–5490.
    57. 57)
      • 12. Zhang, K., Zuo, W., Chen, Y., et al: ‘Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising’, IEEE Trans. Image Process., 2017, 26, (7), pp. 3142–3155.
    58. 58)
      • 2. Healey, G.E., Kondepudy, R.: ‘Radiometric CCD camera calibration and noise estimation’, IEEE Trans. Pattern Anal. Mach. Intell., 1994, 16, (3), pp. 267–276.
    59. 59)
      • 5. Mafi, M., Martin, H., Andrian, J., et al: ‘A comprehensive survey on impulse and Gaussian denoising filters for digital images’, Signal Process., 2019, 157, pp. 236–260.
    60. 60)
      • 52. Xie, J., Xu, L., Chen, E.: ‘Image denoising and inpainting with deep neural networks’. Proc. 25th Int. Conf. Neural Information Processing Systems (NIPS), Lake Tahoe, NV, USA, 2012, pp. 341–349.
    61. 61)
      • 15. Shi, W., Gong, Y., Tao, X., et al: ‘Fine-grained image classification using modified DCNNs trained by cascaded softmax and generalized large-margin losses’, IEEE Trans. Neural Netw. Learn. Syst., 2019, 30, (3), pp. 683–694.
    62. 62)
      • 33. Jiang, J., Zhang, L., Yang, J.: ‘Mixed noise removal by weighted encoding with sparse nonlocal regularization’, IEEE Trans. Image Process., 2014, 23, (6), pp. 2651–2662.
    63. 63)
      • 61. Duchi, J., Hazan, E., Singer, Y.: ‘Adaptive subgradient methods for online learning and stochastic optimization’, J. Mach. Learn. Res., 2011, 12, pp. 2121–2159.
http://iet.metastore.ingenta.com/content/journals/10.1049/iet-ipr.2019.0931
Loading

Related content

content/journals/10.1049/iet-ipr.2019.0931
pub_keyword,iet_inspecKeyword,pub_concept
6
6
Loading