 |
Technical Program
Paper Detail
| Paper: | MP-P4.7 |
| Session: | Biomedical Image Reconstruction |
| Time: | Monday, October 9, 14:20 - 17:00 |
| Presentation: |
Poster
|
| Title: |
BLIND DEBLURRING RECONSTRUCTION TECHNIQUE WITH APPLICATIONS IN SPECT IMAGING |
| Authors: |
Heng Li; University of Texas M.D. Anderson Cancer Center | | | | | Yibin Zheng; University of Virginia | | |
| Abstract: |
To resolve fine details in Single Photon Emission Computed Tomography (SPECT), or to accurately segment tumors from background activities, it is desired for the SPECT reconstruction image to preserve the edges and achieve high spatial resolution. In this paper, we develop a blind deblurring reconstruction technique to jointly estimate both the actual image and the system PSF. A blurred SPECT reconstruction is regarded as the convolution of an unknown low-pass PSF with the true image. Inspired by the blind deconvolution algorithm, we formulate a blind deblurring reconstruction algorithm, which consists of two iterative update equations corresponding to the PSF estimation and the SPECT reconstruction respectively. The PSF of the SPECT system is assumed real, non-negative, and band-limited. In both simulation and experimental studies of piece-wise constant phantoms, the contrast and resolution of the reconstruction is substantially improved. |
|
