基于图像加权的多能光子计数X线CT全能谱图像重建改进方法

2017-10-12 07:45:59周正东管绍林涂佳丽李剑波张雯雯
东南大学学报(自然科学版) 2017年5期

周正东 管绍林 涂佳丽 李剑波 张雯雯

(南京航空航天大学核科学与工程系, 南京 210016)

基于图像加权的多能光子计数X线CT全能谱图像重建改进方法

周正东 管绍林 涂佳丽 李剑波 张雯雯

(南京航空航天大学核科学与工程系, 南京 210016)

为了提高基于图像加权的多能光子计数X线CT全能谱重建图像的质量,提出了一种图像重建改进方法.首先,对每个能量段的投影数据采用最大后验概率(MAP)统计重建算法进行图像重建;然后,对各能量段图像进行优化加权求和,获得全能谱图像.仿真实验结果表明,MAP统计重建算法可显著提高全能谱重建图像的对比噪声比(CNR).与滤波反投影算法相比,当能量段数分别为2,4,6,9时,MAP统计重建算法使得钙图像的CNR分别提高659.7%,643.4%,621.2%,586.1%,碘图像的CNR分别提高663.8%,648.6%,635.1%,600.9%,软组织图像的CNR分别提高596.2%,638.5%,592.6%,596.3%;与能量积分法相比,当能量段数分别为2,4,6,9时,MAP统计重建算法使得钙图像的CNR分别提高43.3%,49.1%,49.3%,44.5%,碘图像的CNR分别提高43.2%,45.7%,45.7%,40.2%,软组织图像的CNR分别提高21.5%,28.9%,25.5%,26.2%.

光子计数X线CT; 基于图像加权; 最大后验概率; 对比噪声比

Abstract: To improve the full spectral reconstruction image quality for the image-based weighting multi-energy photon counting X-ray computed tomography (CT), an improved image reconstruction method was proposed. First, the maximum a posteriori probability (MAP) statistical reconstruction algorithm was used to reconstruct the image in each energy bin. Then, the images from each energy bin were summarized with optimal weights to obtain the full spectral image. The simulation experimental results demonstrate that the MAP statistical reconstruction algorithm can significantly improve the contrast-to-noise ratio(CNR) of the full spectral reconstruction image. Compared with the filtered back projection (FBP) algorithm, for the cases with the energy spectrum split into 2, 4, 6 and 9 bins, the MAP statistical reconstruction algorithm can offer the CNR improvement up to 659.7%, 643.4%, 621.2% and 586.1%for calcium,663.8%, 648.6%, 635.1% and 600.9% for iodine, 596.2%, 638.5%,592.6% and 596.3% for soft tissue, respectively.Compared with the energy-integrating method,for the energy-resolved cases with 2, 4, 6 and 9 energy bins, the MAP statistical reconstruction algorithm can offer the CNR improvement up to 43.3%, 49.1%, 49.3% and 44.5% for calcium, 43.2%, 45.7%, 45.7% and 40.2% for iodine, 21.5%, 28.9%, 25.5% and 26.2% for soft tissue,respectively.

Keywords: photon counting X-ray computed tomography (CT); image-based weighting; maximum a posteriori probability(MAP); contrast-to-noise ratio(CNR)

传统X线CT系统利用电荷积分探测器进行信号检测,光子权重与其能量成正比,较高能量的光子被赋予较大的权重.然而,重建图像的对比度主要取决于较低能量的光子,因此电荷积分探测方式将导致重建图像对比度下降[1-2].近期研发的光子计数探测器具有能量分辨能力,它由半导体探测器及快速专用集成电路组成,在医学诊断X线能量范围内具有较高的量子效率,可灵活设置能……

登录APP查看全文