Evaluation of image filters for their integration with LSQR computerized tomography reconstruction method

[EN] In CT (computerized tomography) imaging reconstruction, the acquired sinograms are usually noisy, so artifacts will appear on the resulting images. Thus, it is necessary to find the adequate filters to combine with reconstruction methods that eliminate the greater amount of noise possible without altering in excess the information that the image contains. The present work is focused on the evaluation of several filtering techniques applied in the elimination of artifacts present in CT sinograms. In particular, we analyze the elimination of Gaussian and Speckle noise. The chosen filtering techniques have been studied using four functions designed to measure the quality of the filtered image and compare it with a reference image. In this way, we determine the ideal parameters to carry out the filtering process on the sinograms, prior to the process of reconstruction of the images. Moreover, we study their application on reconstructed noisy images when using noisy sinograms and finally we select the best filter to combine with an iterative reconstruction method in order to test if it improves the quality of the images. With this, we can determine the feasibility of using the selected filtering method for our CT reconstructions with projections reduction, concluding that the bilateral filter is the filter that behaves best with our images. We will test it when combined with our iterative reconstruction method, which consists on the Least Squares QR method in combination with a regularization technique and an acceleration step, showing how integrating this filter with our reconstruction method improves the quality of the CT images.This research has been supported by "Universitat Politecnica de Valencia", "Generalitat Valenciana" under PROMETEO/2018/035 as well as ACIF/2017/075 predoctoral grant co-financed by FEDER and FSE funds, and "Spanish Ministry of Science, Innovation and Universities" under Grant RTI2018-098156-B-C54 co-financed by FEDER funds.Chillarón-Pérez, M.; Vidal-Gimeno, V.; Verdú Martín, GJ. (2020). Evaluation of image filters for their integration with LSQR computerized tomography reconstruction method. PLoS ONE. 15(3):1-14. https://doi.org/10.1371/journal.pone.0229113114153Managing patient dose in computed tomography. (2000). Annals of the ICRP, 30(4), 7-7. doi:10.1016/s0146-6453(01)00049-5Chillarón, M., Vidal, V., Segrelles, D., Blanquer, I., & Verdú, G. (2017). Combining Grid Computing and Docker Containers for the Study and Parametrization of CT Image Reconstruction Methods. Procedia Computer Science, 108, 1195-1204. doi:10.1016/j.procs.2017.05.065Flores, L., Vidal, V., & Verdú, G. (2015). Iterative Reconstruction from Few-view Projections. Procedia Computer Science, 51, 703-712. doi:10.1016/j.procs.2015.05.188Flores, L. A., Vidal, V., Mayo, P., Rodenas, F., & Verdú, G. (2014). Parallel CT image reconstruction based on GPUs. Radiation Physics and Chemistry, 95, 247-250. doi:10.1016/j.radphyschem.2013.03.011Parcero, E., Flores, L., Sánchez, M. G., Vidal, V., & Verdú, G. (2017). Impact of view reduction in CT on radiation dose for patients. Radiation Physics and Chemistry, 137, 173-175. doi:10.1016/j.radphyschem.2016.01.038I. Kumar, H. Bhadauria, J. Virmani, and J. Rawat, “Reduction of speckle noise from medical images using principal component analysis image fusion,” in Industrial and Information Systems, 2014 9th International Conference on. IEEE, 2014, pp. 1–6.Barrett, J. F., & Keat, N. (2004). Artifacts in CT: Recognition and Avoidance. RadioGraphics, 24(6), 1679-1691. doi:10.1148/rg.246045065Chillarón, M., Vidal, V., Verdú, G., & Arnal, J. (2018). CT Medical Imaging Reconstruction Using Direct Algebraic Methods with Few Projections. Computational Science – ICCS 2018, 334-346. doi:10.1007/978-3-319-93701-4_25Joseph, P. M. (1982). An Improved Algorithm for Reprojecting Rays through Pixel Images. IEEE Transactions on Medical Imaging, 1(3), 192-196. doi:10.1109/tmi.1982.4307572F. P. Group. FORBILD head phantom. [Online]. Available: http://www.imp.uni-erlangen.de/forbild/english/results/index.htm.Paige, C. C., & Saunders, M. A. (1982). LSQR: An Algorithm for Sparse Linear Equations and Sparse Least Squares. ACM Transactions on Mathematical Software, 8(1), 43-71. doi:10.1145/355984.355989Yu, H., & Wang, G. (2010). A soft-threshold filtering approach for reconstruction from a limited number of projections. Physics in Medicine and Biology, 55(13), 3905-3916. doi:10.1088/0031-9155/55/13/022Beck, A., & Teboulle, M. (2009). A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems. SIAM Journal on Imaging Sciences, 2(1), 183-202. doi:10.1137/080716542C. Tomasi and R. Manduchi, “Bilateral filtering for gray and color images,” in Sixth International Conference on Computer Vision. IEEE, 1998, pp. 839–846.A. Hore and D. Ziou, “Image Quality Metrics: PSNR vs. SSIM,” in 2010 20th International Conference on Pattern Recognition. IEEE, aug 2010, pp. 2366–2369