We propose a novel pinhole collimator in which the pinhole shape can be changed in real-time, and a new single-photon emission computed tomography (SPECT) system that utilizes this variable pinhole (VP) collimator. The acceptance angle and distance between the collimator and the object of VP SPECT are varied so that the optimum value of the region-of-interest (ROI) can be obtained for each rotation angle. Because of these geometrical variations, new correction methods are required for image reconstruction. In this study, we developed two correction methods. The first is the sensitivity-correction algorithm, which minimizes the variation of a system matrix caused by varying the acceptance angle for each rotation angle. The second is the acquisition-time-correction method, which reduces the variation of uniformity caused by varying the distance between the collimator and the object for each rotation angle. A 3D maximum likelihood expectation maximization (MLEM) algorithm was applied to image reconstruction, and two digital phantoms were studied to evaluate the resolution and sensitivity of the images obtained using the proposed methods. The images obtained by using the proposed correction methods show higher uniformity and resolution than those obtained without using these methods. In particular, the results of the resolution phantom study show that hot rods (0.8-mm-diameter) can be clearly distinguished using the proposed correction methods. A quantitative analysis of the ROI phantom revealed that the mean square error (MSE) was 0.42 without the acquisition-time-correction method, and 0.04 with the acquisition-time-correction method. The MSEs of the resolution phantom without and with the acquisition-time-correction method were calculated as 55.14 and 14.69, respectively.
Bibliographical noteFunding Information:
This work was supported by the Commercializations Promotion Agency for R&D Outcomes (COMPA) (No. 2014K000268), the National R&D Program through the Korea Institute of Radiological and Medical Sciences (No. 1711021927), and the National Research Foundation of Korea (No. NRF-2015M2B2A9028143) funded by the Ministry of Science, ICT & Future Planning of the Korean government.
© 2016 IOP Publishing Ltd and Sissa Medialab srl.
- Data processing methods
- Gamma camera, SPECT, PET PET/CT, coronary CT angiography (CTA)
- Medical-image reconstruction methods and algorithms, computer-aided software
ASJC Scopus subject areas
- Mathematical Physics