This website requires JavaScript.

Evaluation of low-dose CT supervised learning algorithms with transformer-based model observer

Yongyi ShiGe WangXuanqin Mou
Dec 2022
摘要
A variety of supervise learning methods have been proposed for low-dosecomputed tomography (CT) sinogram domain denoising. Traditional measures ofimage quality have been employed to optimize and evaluate these methods, wheremathematical model observers are widely advocated because it's designed to actas human surrogates in the task of lesion detectability for clinical diagnosis.However, the sinogram domain evaluation for the supervised learning methodsremains lacking. Since the lesion in medical images is correspond to a narrowsine strip in sinogram domain, in this paper, we proposed a transformer-basedefficient model observer to evaluate the sinogram domain supervise learningmethods. The numerical results indicate that, we can well-approximate theLaguerre-Gauss channelized Hotelling observer (LG-CHO) models for asignal-known-exactly (SKE) and background-known-statistically (BKS) task bytraining a transformer-based model observer in sinogram domain. The proposedtransformer-based model observer is then employed to assess three classicCNN-based sinogram domain denoising methods. The evaluation in these denoisingmethods may suggest future avenues for improving the effectiveness of sinogramdomain supervised learning methods.
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