项目名称: 基于分数阶统计建模的低剂量CT优质成像新方法研究
项目编号: No.31470048
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 生物科学
项目作者: 廖志武
作者单位: 四川师范大学
项目金额: 30万元
中文摘要: CT的高辐射大大增加了人体罹患癌症的风险。但是CT剂量的降低会导致重建图像产生严重的噪声和伪影。近年来,低剂量CT成像校正算法已经成为医学成像领域的热点问题。由于低剂量CT投影数据、重建图像和噪声统计特性非常复杂,难以用整数阶统计分布和统计量进行描述,因此本项目拟利用分数阶建模拟合低剂量CT数据的复杂统计特性,在此基础上,拟开展以下几方面的研究:(1)分数阶目标函数构建,研究如何引入分数阶先验,并在此基础上构造有效的分数阶混合多先验目标函数;(2)分数阶偏微分方程在投影数据校正和重建图像后处理校正中的研究。
中文关键词: 多先验;分数阶;稀疏表示;边缘先验;堆叠2D矩阵
英文摘要: The high radiation of CT will increase the risk of cancers. However, the reduced dose of CT will leads to serious noises and artifacts in reconstructed images. Therefore, how to improve the performance of low dose CT imaging becomes a hot topic in the medical imaging. But the statistical properities of sinograms, reconstructed images and their noises of low dose CT are very complex, which can not be fitted by integer order statistical distributions and amounts. In this project, we will model sinograms, reconstructed images and their noises by fractional order statistics, which is called fractional order statistical modeling (FOSM). Based on FOSM, we will carry on the follwing studies: (1) designing the target functions based on FOSM, which includes the target functions of sinogrm correction models, iterated reconstruction correction models and reconstructed image correction models. The key problem is how to designing target functions based fractional order mixed multi priors (FOMMPs); (2) studying the theories of adaptively and quickly searching algorithms of optimal resolutions based on fractional order variation theory; (3) studying correction algorithms of sinograms and reconstructed images using fractional order partial differential equations (FOPDEs); (4) modeling between blind evaluations of clinic doctor
英文关键词: multi-prior;fractional-order;sparse representation;edge proir;stacked 2D matrixes