Three-dimensional restoration of single photon emission computed tomography images

Abstract

Restoration of single photon emission computed tomography (SPECT) images is usually performed either on two-dimensional (2-D) projection images (pre-reconstruction restoration) or on SPECT slices (post-reconstruction restoration). This study presents the application of a three-dimensional (3-D) filter in restoring SPECT images. The 3-D filter is shown to perform better than the 2-D methods because it takes into consideration the interslice information in the filtering process and has a relatively space-invariant blur function. To reduce the space-variance of the 3-D point spread function (PSF), conjugate projections were combined by geometric averaging before reconstruction. We investigated the potential of 3-D Wiener and power spectrum equalization (PSE) filters. These filters were applied to SPECT images of a resolution phantom and a large, truncated cone phantom containing two types of cold spots: a sphere, and a triangular prism. The filters were also applied to clinical images of the liver and spleen. The images were acquired on an ADAC GENESYS camera. A comparison was performed between results obtained in this study and those obtained by 2-D pre-reconstruction restoration and 2-D post-reconstruction restoration filters. Quantitative analysis of the 3-D restored images performed through measurement of root mean squared (RMS) errors and contrast ratios showed a considerable reduction in error and increase in contrast over images restored using the two other methods. >