http://www.paper.edu.cn
- 1 -
中国科技论文在线
Three-image-based multiple information retrieval algorithm
for X-ray Grating-based Imaging
#
JIANG Xiaolei, HUANG Zhifeng
**
(1. Department of Engineering Physics, Tsinghua University, Beijing, 10084, China) 5
Foundations: The work was supported by a grant from the National Natural Science Foundation of China (NNSFC)
grants 10905031 and 10875066 and the Specialized Research Fund for the Doctoral Program of Higher Education
grant 20090002120016.
Brief author introduction:JIANG Xiaolei, male, Master Candidate, X-ray grating-based imaging
Correspondance author: HUANG Zhifeng, male, assistant professor, X-ray grating-based imaging. E-mail:
huangzhifeng@mail.tsinghua.edu.cn
Abstract: Hard X-ray grating-based imaging has the most potential in clinical applications because it
works well with conventional X-ray tubes and provides multiple information (absorption, refraction
and scattering) simultaneously. However, grating-based imaging adopts a series of images to retrieve
multiple information, which significantly increases the exposure time and doses compared with
conventional X-ray imaging. In this paper, we propose an optimized three-image-based multiple
10
information retrieval (TIB-MIR) algorithm for low-dose grating-based imaging. The errors of the
retrieval information are minimal in the case of three steps with equal intervals of within one period,
which is equivalent to the 3-step fringe-scanning algorithm. Furthermore, in the case of Poisson noise
model and equal doses, the method can obtain equivalent image quality compared with fringe-scanning
method.
15
Key words: Phase contrast imaging, fringe-scanning, Poisson noise model
0 Introduction
Hard X-ray phase contrast imaging has developed rapidly recently. Among various methods,
grating-based imaging (GBI) [1-4] has the most potential for clinical applications because of
its compatibility of conventional X-ray tubes. However, a series of images are needed to 20
retrieve multiple information by use of the phase-stepping approach, which significantly
increase the imaging time and doses.
To solve this problem, the three-image-based multiple information retrieve(TIB-MIR)
algorithm has been proposed[5]. In this paper, the algorithm is optimized based on the Poisson
noise model. The errors of the retrieval information are minimal in the case of three steps with 25
equal intervals of
2/3
within one period, which is equivalent to the 3-step fringe scanning
algorithm. Furthermore, in the case of equal total doses, the method can obtain equivalent
image quality compared with N-step fringe-scanning method, which can significantly
decrease imaging time without loss of image quality.
1 Methods 30
1.1 Optimized TIB-MIR algorithm
The detailed description of the GBI method can be found in Ref.[2,4]. The phase-stepping
approach captures a series of raw images, in which the source grating is fixed and the last two
gratings are relatively moved over one grating period [1], or the last two gratings are fixed and the
source grating is moved only [5]. The sample intensity oscillation curve of each pixel on the 35
detector are measured by the phase stepping approach, which can be expressed by
()
cos 2
B
k
Ik a b
p
χ
ϕ
⎛⎞
≈+ +
⎜⎟
⎝⎠
(1)