.png){kind=icon}
.png){kind=icon}
Return
A Simple, Dynamic Geometric Phantom for MRI and CT Reconstruction Pipelines: Beyond Shepp–Logan
.png){kind=icon}
OA
AI
T
N
J
D
K
DOI：10.1002/mrm.70495.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A customizable, dynamic geometric phantom was developed to bridge the gap between simplistic static phantoms and highly realistic, complex 4D solutions. The phantom enables initial testing, helps detect implementation errors, and supports quantitative evaluation of new dynamic acquisition and reconstruction methods. It is formed from a series of superellipsoids using the constructive solid geometry technique and serves as a schematic representation of a human torso, including the lungs, liver, and stomach, as well as a four-chamber heart, major vessels, and simplified skeletal structures. This design allows full parameterization of voxel intensities and organ volumes. Respiratory and cardiac motions are modeled using built-in, decoupled signal generators. A high-performance, open-source reference implementation was developed in Julia and is available both as a library and a standalone console application. The phantom's capabilities were demonstrated through four experiments. The difference between the input and measured lung volumes was quantified, yielding a maximum relative error of 5.5% and a median error of 2.5%. The relative error of the cardiac chamber volumes ranged from 1.76% to 3.67%. Performance was benchmarked across multiple phantom sizes and thread counts, and a realistic Cartesian MRI simulation demonstrated how the phantom can be used to quantify motion blur and artifacts. The proposed dynamic geometric phantom provides an intermediate option between simple static phantoms and complex, proprietary, anatomically realistic 4D models. It provides a highly customizable, easy-to-use, open-source tool for researchers developing new reconstruction algorithms, and it can be easily integrated into existing reconstruction pipelines.
Keywords:
constructive solid geometry
CT
dynamic phantom
Julia
MRI
Shepp–Logan
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3
Papers:
1.2W
Citations:
3.1W
Organization
Cited Papers