Andrey Krylov
ABSTRACT
Convolutional neural networks (CNN) show very good performance and achieve impressive results in biomedical image denoising. Nevertheless CNN-based methods strongly depend on the used training set and even small differences in the input data can cause output disturbance. Thus new more reliable hybrid denoising methods were suggested. They include combinations of CNN and "classical" algorithms like Non-Local Means, BM3D, Bilateral, Anisotropic diffusion, Total Variation (TV), etc. However hybrid methods need non-reference automatic parameters estimation for classical algorithms. In this paper we present a hybrid DnCNN+BM3D method with automatic choice of the strength parameter for BM3D method. To control biomedical image structures by multiscale ridge based approach we analyze presence of regular structures in the ridge areas at the difference between noisy and filtered images. Test results for retinal image dataset DRIVE, natural images BSD and for real CT image show practical applicability of the method.
- P. Kaur, G. Singh, and P. Kaur. 2018. A Review of Denoising Medical Images Using Machine Learning Approaches. Current medical imaging reviews, 14(5). 675--685. DOI= https://doi.org/10.2174/1573405613666170428154156Google Scholar
- S. Renukalatha, and K. V. Suresh. 2018. A review on biomedical image analysis. Biomedical Engineering: Applications, Basis and Communications, 30(04). 1830001. DOI= https://doi.org/10.4015/S1016237218300018Google ScholarCross Ref
- M. Diwakar, and M. Kumar. 2018. A review on CT image noise and its denoising. Biomedical Signal Processing and Control, 42. 73--88. DOI= https://doi.org/10.1016/j.bspc.2018.01.010Google Scholar
- L. L. Diaz, A. C.Ol mos, A. Cadena, W. S. Percybrooks, and P. J. P. Tello. 2018. Three-dimensional reconstruction of cardiac structures using CT images. In Proceedings of the 2018 3rd International Conference on Biomedical Imaging, Signal Processing, 74--79. DOI= https://doi.org/10.1145/3288200.3288210Google ScholarDigital Library
- N. Obukhova, A. Motyko, A. Pozdeev, and B. Timofeev. 2018. Review of Noise Reduction Methods and Estimation of their Effectiveness for Medical Endoscopic Images Processing. In 2018 22nd Conference of Open Innovations Association (FRUCT), 204--210. DOI= https://doi.org/10.23919/FRUCT.2018.8468285Google ScholarCross Ref
- T. Joel, and R. Sivakumar. 2018. An extensive review on Despeckling of medical ultrasound images using various transformation techniques. Applied Acoustics, 138. 18--27. DOI= https://doi.org/10.1016/j.apacoust.2018.03.023Google ScholarCross Ref
- K. Zhang, W. Zuo, Y. Chen, D. Meng, and L. Zhang. 2017. Beyond a Gaussian denoiser: Residual learning of deep CNN for image denoising. IEEE Trans. Image Process., 26(7), 3142--3155. DOI= https://doi.org/10.1109/TIP.2017.2662206Google ScholarDigital Library
- D. Ulyanov, A. Vedaldi, and V. Lempitsky. 2018. Deep image prior. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 9446--9454.Google Scholar
- D. Liu, B. Wen, Y. Fan, C. C. Loy, and T. S. Huang. 2018. Non-local recurrent network for image restoration. In Advances in Neural Information Processing Systems. 1673--1682.Google Scholar
- C. Cruz, A. Foi, V. Katkovnik, and K. Egiazarian. 2018. Nonlocality-reinforced convolutional neural networks for image denoising. IEEE Signal Processing Letters, 25(8). 1216--1220. DOI= https://doi.org/10.1109/LSP.2018.2850222Google ScholarCross Ref
- G. Mataev, M. Elad, and P. Milanfar. 2019. DeepRED: Deep Image Prior Powered by RED. arXiv preprint, arXiv:1903.10176.Google Scholar
- J. Liu, Y. Sun, X. Xu, and U. A. Kamilov. 2018. Image restoration using total variation regularized deep image prior. arXiv preprint, arXiv:1810.12864.Google Scholar
- N. V. Mamaev, D. V. Yurin, and A. S. Krylov. 2018. Finding the parameters of a nonlinear diffusion denoising method by ridge analysis. Computational Mathematics and Modeling 29(3). 334--343. DOI= https://doi.org/10.1007/s10598-018-9413-6Google ScholarDigital Library
- N. Mamaev, D. Yurin, and A. Krylov. 2018. Choice of the parameter for BM3D denoising algorithm using no- reference metric. In 2018 7th European Workshop on Visual Information Processing (EUVIP). 1--6. DOI= https://doi.org/10.1109/EUVIP.2018.8611679Google ScholarCross Ref
- A. Dovganich, N. Mamaev, A. Krylov., and N. Makhneva. 2019. Dermatological image denoising using adaptive HeNLM method. ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 42(2/W12). 47--52. DOI= https://doi.org/10.5194/isprs-archives-xlii-2-w12-47-2019Google Scholar
- K. Dabov, A. Foi, V. Katkovnik, and K. Egiazarian. 2007. Image denoising by sparse 3-D transform-domain collaborative filtering. IEEE Transactions on Image Processing, 16(8). 2080-2095. DOI= https://doi.org/10.1109/TIP.2007.901238Google ScholarDigital Library
- https://github.com/iorism/CNN/tree/master/testsets, accessed: 2019-05-31.Google Scholar
- J. Staal, M. D. Abràmoff, M. Niemeijer, M. A. Viergever, and B. Van Ginneken. 2004. Ridge-based vessel segmentation in color images of the retina. IEEE transactions on medical imaging, 23(4). 501--509. DOI= https://doi.org/10.1109/TMI.2004.825627Google Scholar
Index Terms
- Hybrid Method for Biomedical Image Denoising
Recommendations
Hybrid Method for Biomedical Image Poisson Denoising
ICBIP '20: Proceedings of the 5th International Conference on Biomedical Signal and Image ProcessingConvolutional neural networks (CNN) have recently become the main tool for the problem of biomedical image denoising. Nevertheless, CNN-based methods strongly depend on the used training set, and even small differences in the input data can cause output ...
Improved KMV-Cast with BM3D Denoising
Recently, a novel scheme called Knowledge-enhanced Mobile Video Broadcasting (KMV-Cast) was proposed to improve the quality of mobile video transmission. In this paper, we propose to use the well-known block-matching and three dimensional filtering (...
A New Adaptive TV-Based BM3D Algorithm for Image Denoising
Artificial IntelligenceAbstractBlock matching 3D filtering (BM3D) algorithm is more effective than traditional denoising methods especially for Gaussian noise. However, the traditional hard-threshold used in BM3D algorithm can not recognize the noise intensity in the process of ...
Comments