2.5D Extension of Neighborhood Filters for Noise Reduction in 3D Medical CT Images
- 610 Downloads
Noise in 3D computer tomography (CT) images is close to white and becomes large when patient radiation doses are reduced. We propose two methods for noise reduction in CT images: 3D extension of fast rank algorithms (Rank-2.5D) and 3D extension of a non-local means algorithm (NLM-2.5D). We call both our algorithms “2.5D” because the extended NLM algorithm is slightly asymmetric by slice axes, while our Rank algorithms, being fully symmetric mathematically and by results, have some implementation asymmetry. A comparison of the methods is presented. It is shown that NLM-2.5D method has the best quality, but is computationally expensive: its complexity quickly rises as a function of the neighborhood size, while Rank-2.5D only shows linear growth. Another contribution of this paper is a modified multiscale histogram representation in memory with a tree-like structure. This dramatically reduces memory requirements and makes it possible to process 16-bit DICOM data with full accuracy. Artificial test sequences are used for signal-to-noise performance measurements, while real CT scans are used for visual assessment of results. We also propose two new measures for no-reference denoising quality assessment based on the autocorrelation coefficient and entropy: both measures analyze randomness of the difference between noisy and filtered images.
Keywordsmedical imaging CT DICOM filtering enhancement noise reduction denoising 3D image processing image quality assessment
Unable to display preview. Download preview PDF.
- 1.Pratt, W.: Digital Image Processing: PIKS Scientific Inside. Wiley (2007)Google Scholar
- 2.Yaroslavsky, L., Kim, V.: Rank algorithms for picture processing. ACM Transactions on Graphics (TOG) 35, 234–258 (1986)Google Scholar
- 6.Storozhilova, M., Yurin, D.: Fast rank algorithms based on multiscale histograms. In: 21st International Conference on Computer Graphics GraphiCon 2011, Moscow, Russia, pp. 132–135 (September 2011)Google Scholar
- 7.Storozhilova, M., Yurin, D.: Fast rank algorithms with multiscale histograms lazy updating. In: 8th Open German-Russian Workshop “Pattern Recognition and Image Understanding” (OGRW-8-2011), Lobachevsky State University of Nizhny Novgorod, pp. 380–383 (November 2011)Google Scholar
- 8.Tomasi, C., Manduchi, R.: Bilateral filtering for gray and color images. In: Proceedings of the IEEE Sixth International Conference on Computer Vision (ICCV 1998), pp. 839–846 (1998)Google Scholar
- 9.Buades, A., Morel, J.: A non-local algorithm for image denoising. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2, pp. 60–65 (2005)Google Scholar
- 11.Reiter, M., Zauner, G.: Denoising of computed tomography images using multiresolution based methods. In: European Conference on Non-Destructive Testing poster (2006)Google Scholar
- 12.Trinh, D., Luong, M., Rocchisani, J.M., Pham, C., Pham, H., Dibos, F.: An optimal weight method for CT image denoising. Journal of Electronic Science and Technology 10(2), 124–129 (2012)Google Scholar
- 14.Newton, T., Potts, D.: Radiology of the Skull and Brain: Technical Aspects of Computed Tomography. Radiology of the Skull and Brain. Mosby (1981)Google Scholar
- 15.Putilin, S., Lukin, A.: Non-local means method modification for noise suppression in video. In: 17th International Conference on Computer Graphics, GraphiCon 2007, pp. 257–259 (June 2007) (in Russian)Google Scholar
- 16.Storozhilova, M., Lukin, A., Yurin, D., Sinitsyn, V.: Two approaches for noise filtering in 3D medical CT-images. In: 22nd International Conference on Computer Graphics, GraphiCon 2012, Moscow, Russia, pp. 68–72 (October 2012)Google Scholar