Figure 2
First test case. (a) Original phantom. (b) Result of filtered back-projection after adding constant lines in the sinogram. (c) Image back-projected after applying the Munch et al. de-striper algorithm with = 3.5, L = 2 and the `Daubechies 15' wavelet. (d) Difference between the phantom and the corrected image. The PSNR is 26.6. (e) Result of the correction with the RCP technique with W = 10 and = 10. (f) Difference between the phantom and the corrected image. The PSNR is 29.6. (g) Result of the reconstruction using the Total Variation regularization, with parameters = 0.5, = 0.05. (h) Difference between the phantom and the corrected image. The PSNR is 39.0. (i) Result of the reconstruction using the Dictionary Learning technique with = 0.3, = 0.5, = 1. (j) Difference between the phantom and the reconstructed image. The PSNR is 29.2. |