Fast implementation and quantitative evaluation of analytical methods withwiener filters for image reconstruction in 3D SPECT

We implemented analytical methods for image reconstruction in three-dimensi onal single-photon emission computed tomography with uniform attenuation an d distance-dependent blurring. The implementation is computationally effici ent and requires less than 4 minutes for reconstruction of a 128 x 128 x 12 8 image on a personal computer. We conducted a quantitative evaluation of t he reconstruction methods and showed that, when deblurring is considered, t hese reconstruction methods are highly susceptible to noise. We subsequentl y developed a Wiener apodizing window function for suppression of noise amp lification in the reconstructed images. We also evaluated the performance o f the proposed Wiener window function and the frequently used truncation wi ndow function with both noiseless and noisy data. Our numerical results dem onstrated that the images generated from noisy data by use of the proposed Wiener window function have fewer ring-like artifacts and comparable quanti tative accuracy in comparison with those obtained by use of the truncation window function.