The multi-source instant CT for superfast imaging: system concept, reconstruction algorithms and experiments

Abstract

In recent years, the dynamic CT technology has attracted more and more attentions in medical and industrial applications. Instead of increasing the rotation speed of the gantry because of the mechanical limits, a more feasible method is to increase the amount of x-ray sources and detectors like the dual-source CT of Siemens. This paper focuses high resolution CT imaging (e.g. 50µm) of in vivo human body. We propose and try to develop a new CT system named instant CT which uses about 35 couples of carbon nanotube (CNT) x-ray sources and flat-panel detectors. All of these CNT x-ray sources and detectors are installed on the same gantry. It scans the target region-of-interest (ROI) dozens of times to collect the sufficient amount of projection data. At each exposure position, all of the x-ray sources are fast pulse-exposed synchronously, and the flat-panel detectors collect the projection data. Then, the gantry rotates a little angle to prepare for the next exposure until the rotation covers the angle between two x-ray sources. Because each exposure is very fast (about 0.1∼5ms) the organ motions of in vivo human body can be greatly reduced. This paper introduces the instant CT system design, image reconstruction algorithms and experimental results. The image reconstruction for this multi-source instant CT includes three steps: few-view ROI image reconstruction from one simultaneous exposure data, motion registration among the sequence CT images from different exposures, and high resolution ROI image reconstruction with motion correction from all of the sequence projection data. Experiments were designed on our micro-CT system with a homemade phantom. The results validated the efficiency of this new CT system and reconstruction algorithms.