This website requires JavaScript.
DOI: 10.1007/978-3-031-16446-0_63

Orientation-Shared Convolution Representation for CT Metal Artifact Learning

Hong WangQi XieYuexiang LiYawen HuangDeyu MengYefeng Zheng
Dec 2022
摘要
During X-ray computed tomography (CT) scanning, metallic implants carryingwith patients often lead to adverse artifacts in the captured CT images andthen impair the clinical treatment. Against this metal artifact reduction (MAR)task, the existing deep-learning-based methods have gained promisingreconstruction performance. Nevertheless, there is still some room for furtherimprovement of MAR performance and generalization ability, since some importantprior knowledge underlying this specific task has not been fully exploited.Hereby, in this paper, we carefully analyze the characteristics of metalartifacts and propose an orientation-shared convolution representation strategyto adapt the physical prior structures of artifacts, i.e., rotationallysymmetrical streaking patterns. The proposed method rationally adoptsFourier-series-expansion-based filter parametrization in artifact modeling,which can better separate artifacts from anatomical tissues and boost the modelgeneralizability. Comprehensive experiments executed on synthesized andclinical datasets show the superiority of our method in detail preservationbeyond the current representative MAR methods. Code will be available at\url{https://github.com/hongwang01/OSCNet}
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答