基于圖像加權的多能光子計數X線CT全能譜圖像重建改進方法
發(fā)布時間：2018-11-11 16:47
【摘要】：為了提高基于圖像加權的多能光子計數X線CT全能譜重建圖像的質量,提出了一種圖像重建改進方法.首先,對每個能量段的投影數據采用最大后驗概率(MAP)統(tǒng)計重建算法進行圖像重建;然后,對各能量段圖像進行優(yōu)化加權求和,獲得全能譜圖像.仿真實驗結果表明,MAP統(tǒng)計重建算法可顯著提高全能譜重建圖像的對比噪聲比(CNR).與濾波反投影算法相比,當能量段數分別為2,4,6,9時,MAP統(tǒng)計重建算法使得鈣圖像的CNR分別提高659.7%,643.4%,621.2%,586.1%,碘圖像的CNR分別提高663.8%,648.6%,635.1%,600.9%,軟組織圖像的CNR分別提高596.2%,638.5%,592.6%,596.3%;與能量積分法相比,當能量段數分別為2,4,6,9時,M AP統(tǒng)計重建算法使得鈣圖像的CNR分別提高43.3%,49.1%,49.3%,44.5%,碘圖像的CNR分別提高43.2%,45.7%,45.7%,40.2%,軟組織圖像的CNR分別提高21.5%,28.9%,25.5%,26.2%.
[Abstract]:In order to improve the quality of image reconstruction based on image weighted multi-energy photon counting X-ray CT omnipotent spectrum, an improved image reconstruction method is proposed. Firstly, the maximum posterior probability (MAP) statistical reconstruction algorithm is used to reconstruct the projection data of each energy segment, and then the weighted sum of each energy segment image is optimized to obtain the omnipotent spectral image. The simulation results show that the MAP statistical reconstruction algorithm can significantly improve the contrast noise ratio of the reconstructed image with full energy spectrum (CNR). Compared with the filtered backprojection algorithm, when the number of the energy segments is 2? 4? The CNR of soft tissue images increased 596.2 and 638.5and 592.6and 596.3respectively. Compared with the energy integration method, the CNR of calcium image was increased by 43.3% and 49.3%, respectively, when the number of energy segments was 2 ~ 4 ~ 6 and 9, M AP, respectively, and the CNR of iodine image was increased by 43.2 and 44.5, respectively, compared with the energy integration method, and the CNR of iodine image was increased by 43.2%, respectively. The CNR of the soft tissue images increased by 21.5and 28.9and 25.5and 26.2cm, respectively.
【作者單位】： 南京航空航天大學核科學與工程系;
【基金】：國家自然科學基金資助項目(51575256) 中央高�；究蒲袠I(yè)務費專項資金資助項目(NZ2016102)
【分類號】：TP391.41
[Abstract]:In order to improve the quality of image reconstruction based on image weighted multi-energy photon counting X-ray CT omnipotent spectrum, an improved image reconstruction method is proposed. Firstly, the maximum posterior probability (MAP) statistical reconstruction algorithm is used to reconstruct the projection data of each energy segment, and then the weighted sum of each energy segment image is optimized to obtain the omnipotent spectral image. The simulation results show that the MAP statistical reconstruction algorithm can significantly improve the contrast noise ratio of the reconstructed image with full energy spectrum (CNR). Compared with the filtered backprojection algorithm, when the number of the energy segments is 2? 4? The CNR of soft tissue images increased 596.2 and 638.5and 592.6and 596.3respectively. Compared with the energy integration method, the CNR of calcium image was increased by 43.3% and 49.3%, respectively, when the number of energy segments was 2 ~ 4 ~ 6 and 9, M AP, respectively, and the CNR of iodine image was increased by 43.2 and 44.5, respectively, compared with the energy integration method, and the CNR of iodine image was increased by 43.2%, respectively. The CNR of the soft tissue images increased by 21.5and 28.9and 25.5and 26.2cm, respectively.
【作者單位】： 南京航空航天大學核科學與工程系;
【基金】：國家自然科學基金資助項目(51575256) 中央高�；究蒲袠I(yè)務費專項資金資助項目(NZ2016102)
【分類號】：TP391.41
【相似文獻】
相關期刊論文 前10條
1 孫鳳文,郭海峰,楊康;求解光學CT圖像重建問題的擾動方法[J];沈陽工業(yè)學院學報;2002年01期
2 郭海峰,楊康;求解光學CT圖像重建問題的最佳攝動量法[J];儀器儀表學報;2001年S2期
3 潘晉孝,韓焱,繆麗娜;圖像重建中濾波算子的一種優(yōu)化方法[J];測試技術學報;2001年04期
4 夏平;在影像專業(yè)講授“圖像重建”的教學實踐[J];電氣電子教學學報;2001年03期
5 張連順,張春平,王新宇,祁勝文,田建國;光與生物組織相互作用模型及圖像重建[J];激光與光電子學進展;2002年04期
6 海鷹,姜紅軍,平子良;圓諧-傅立葉矩圖像重建誤差分析[J];內蒙古師范大學學報(自然科學漢文版);2003年04期
7 肖斌;ICT圖像重建并行處理技術簡介[J];計量技術;2004年09期
8 張兆田,張朋;改進的圖像重建迭代算法[J];電子與信息學報;2004年10期
9 任變青,潘晉孝;嶺估計算法作圖像重建的仿真研究[J];華北工學院學報;2004年06期
10 潘晉孝,韓焱,劉清s，
本文編號：2325498
本文鏈接：http://www.lk138.cn/kejilunwen/ruanjiangongchenglunwen/2325498.html
