Speaker
Prof.
Hongbing Lu
(Department of Computer Application, Fourth Military Medical University, Shaanxi, China)
Description
Dynamic or gated cardiac SPECT imaging has been playing a major role in evaluation
of chest pain and other cardiac symptoms. Conventionally, the acquired dynamic data
sequences are reconstructed frame by frame, where the temporal correlation among
different frames is not utilized. In this work, we proposed a framework for
analytical reconstruction in the Karhunen-Loeve (KL) domain of quantitative dynamic
SPECT with both noise reduction and accurate compensation for non-uniform
attenuation. The dynamic data sequences were first transformed into the KL domain,
in which the spatio-temporal reconstruction of the entire sequences becomes a series
of reconstruction on the independent KL components. To suppress noise propagation
through the reconstruction, a penalized weighted least-squares (PWLS) filtering was
applied and then the filtered KL components were reconstructed one-by-one based on
the Novikov’s inverse formula which compensates for non-uniform attenuation. By
the proposed framework, a filtered back-projection (FBP)-type inversion for
quantitative dynamic SPECT is feasible with accurate compensation for non-uniform
attenuation and effective treatment of Poisson noise. Our computer simulation
results are very encouraging, by both qualitative and quantitative judgments, as
compared to both the frame-by-frame Novikov inversion and the conventional FBP
reconstruction.
Authors
Prof.
Hongbing Lu
(Department of Computer Application, Fourth Military Medical University, Shaanxi, China)
Mr
Yi Fan
(Department of Electronic Information, North-Western Polytechnological University, Shaanxi, China)
Co-authors
Prof.
Chongyang Hao
(Department of Electronic Information, North-Western Polytechnological University, Shaanxi , China)
Prof.
Zhengrong Liang
(Department of Radiology, State University of New York, Stony Brook, NY, USA)
