發(fā)布時(shí)間：2018-02-16 19:48

  本文關(guān)鍵詞： 錐束CT 最大后驗(yàn)概率 MRF 非局部 有序子窗搜索　出處：《東南大學(xué)學(xué)報(bào)(自然科學(xué)版)》2017年05期 　論文類(lèi)型：期刊論文

【摘要】：為了在稀疏角度掃描條件下更好地去除重建圖像中的條狀偽影和保留細(xì)節(jié)信息,將非局部先驗(yàn)引入錐束CT重建.基于有序子集投影劃分思想,提出了有序子窗搜索算法,用以解決錐束CT迭代重建算法中非局部先驗(yàn)計(jì)算量過(guò)大的問(wèn)題.該算法將每一個(gè)體素的搜索窗劃分為M個(gè)不重復(fù)的子窗,每次迭代中選取不同子集元素計(jì)算非局部先驗(yàn)約束.實(shí)驗(yàn)結(jié)果表明,通過(guò)非局部先驗(yàn)約束,可以獲得質(zhì)量更好的重建圖像.而且無(wú)論是在主觀視覺(jué)效果方面,還是在峰值信噪比和結(jié)構(gòu)相似性指標(biāo)等客觀評(píng)價(jià)指標(biāo)方面,有序子窗搜索算法和傳統(tǒng)非局部算法的重建結(jié)果均無(wú)明顯差別,但前者可以明顯降低先驗(yàn)項(xiàng)的時(shí)間復(fù)雜度.
[Abstract]:In order to remove the strip artifacts and preserve the details in the reconstructed images better under sparse angle scanning, the nonlocal priori is introduced into the cone-beam CT reconstruction. Based on the idea of projection partition of ordered subsets, an ordered sub-window search algorithm is proposed. The algorithm is used to solve the problem that the non-local prior computation of the iterative reconstruction algorithm of cone beam CT is too large. The search window of each individual prime is divided into M non-repeated sub-windows. In each iteration, different subset elements are selected to calculate the nonlocal priori constraints. The experimental results show that the reconstruction images with better quality can be obtained by the nonlocal priori constraints. There is no significant difference in the reconstruction results between the ordered sub-window search algorithm and the traditional nonlocal algorithm in terms of the objective evaluation indexes such as peak signal-to-noise ratio (PSNR) and structural similarity index, but the former can significantly reduce the time complexity of the priori term.
【作者單位】： 東南大學(xué)計(jì)算機(jī)科學(xué)與工程學(xué)院;東南大學(xué)計(jì)算機(jī)網(wǎng)絡(luò)和信息集成教育部重點(diǎn)實(shí)驗(yàn)室;東南大學(xué)附屬中大醫(yī)院血管外科;
【基金】：國(guó)家自然科學(xué)基金資助項(xiàng)目(81530060) 東南大學(xué)計(jì)算機(jī)網(wǎng)絡(luò)和信息集成教育部重點(diǎn)實(shí)驗(yàn)室開(kāi)放課題資助項(xiàng)目(K93-9-2016-07)
【分類(lèi)號(hào)】：R814.42;TP391.41
，

本文編號(hào)：1516301