 |
Z. Soldatova, A. Ingacheva, M. Chukalina, S. Chervonookaya, K. Bulatov Towards monitored tomographic reconstruction: the comparison of algorithms convergence on synthetic data |
 |
|
Abstract.
Monitored tomographic reconstruction (MTR) is a novel approach of dose reduction in computed tomography (CT), which also allows to reduce experiment time in micro tomography setups. The core of MTR is simultaneous projection data registration and its reconstruction with the stop of acquisition process after some predefined quality has been achieved. The order of projection acquisition is defined in the scanning protocol by the means of calibration on the test objects of a similar structure and under similar experimental conditions. The MTR is able to achieve a fixed average reconstruction quality with lower average dose/projection count than the conventional CT protocols allows. In the present work we study MTR on the synthetic data for three different reconstruction algorithms (FBP, SIRT, SIRT-TV) and three different image quality metrics. We found that even though the optimal projections count under predefined quality is algorithm and metric dependent, the general idea still holds, and the gain of MTR over conventional protocol can be achieved.
Keywords:
computed tomography, monitored tomography reconstruction, reconstruction algorithm, quality function, dose reduction, stopping rule.
PP. 10-18.
DOI: 10.14357/20790279220302
|
