基于连续X射线谱的CT重建算法仿真

关于提高CT图像精度的问题,传统的CT重建算法都是基于X射线源是单色源的假设,忽略了X射线的多色性.直接用多色投影数据进行图像重建易产生金属、硬化等伪影,降低图像质量,影响CT值标定,从而影响医学或工业诊断.考虑到X射线能谱的连续性,采用仿真手段实现连续X射线谱的统计重建.首先将连续X射线谱离散成若干单能谱,再根据待检工件的材质信息以及射线能量所对应的质量衰减系数,构建基于连续X射线谱的工件材质模型;最后利用多能统计重建算法对多能投影数据进行迭代重建.仿真结果表明,算法充分地利用了X射线的多能性,在一定程度上可以有效地降低图像伪影,提高CT重建图像质量.     

基于NL-PF和MIMS的CT金属伪影消除算法

建立了一套针对由金属伪影造成的CT图像质量退化的恢复算法。利用Non-Local前置滤波(Non-Local Pre-filter,NL-PF)对原始CT图像进行全局滤波,从而有效地滤除原始图像中的噪声并对射线状金属伪影进行了平滑,其后配合最大互信息量分割算法(Mutual Information Maximized Segmentation,MIMS)从图像中分割出伪影成份,并利用其周围非伪影部分的像素对伪影类像素进行插值处理得到一个称之为"伪组织"类的图像。最后,通过融合"伪组织"图像的sinogram和原始CT图像的sinogram,得到校正的sinogram并采用滤波反投影重建算法完成金属伪影的CT校正图像。利用所提出的方法可以对含有金属伪影的CT图像进行有效伪影消除,其中射线状伪影消除效果显著。另外,此方法还可以锐化器官轮廓,避免了临床上由于金属伪影导致的放射治疗效果下降。实验表明,金属伪影消除算法可以有效地消除高密度物体造成的金属伪影,从而提高临床诊断和治疗的效果提供技术支持。     

基于数学形态学的CT图像金属伪影消除算法

金属伪影的有效校正一直都是CT重建的一个重要课题。传统的插值校正法和迭代去模糊法将投影数据作为输入值,但在实际应用中经常不能得到原始投影数据。针对该问题,提出一种基于数学形态学的金属伪影消除算法。首先,通过线检测和霍夫变换提取初始图像中的条纹伪影,并找到金属物体的中心坐标;其后,以金属物体的中心坐标为注视点,对含有伪影的初始图像进行极坐标变换;继而在极坐标图像中,选择合适的结构元素和形态学过滤器消除伪影;最后将过滤后的极坐标图像转换为笛卡尔坐标系中的图像。实验结果表明,该算法能够有效消除金属伪影,并将滤波反投影图像作为输入值,能够应用于无法得到原始投影数据的情况。     

锥束螺旋CT的串扰屏蔽

在锥束CT(Computed Tomography)中,由于X射线特别是高能X射线散射引起的探测器通道间的串扰,会降低探测器的信噪比和物体的空间分辨率,同时也会在图像的边缘产生伪影.采用了一种有利于机械实现的交错螺旋和平行螺旋的扫描方法,通过在探测器面板上加隔离网来减少射线串扰,并采集4次投影数据,以弥补加隔离网后投影数据的缺损.同时对改进的扫描方法和投影采集方法作了计算机仿真,从仿真实验的结果来看,采用探测器4次采样和交错螺旋或平行螺旋的扫描轨迹,能有效减少因射线串扰引起的伪影,并减少了由螺旋运动引起的伪影.     

基于双调和插值的锥束CT金属伪影校正算法

在计算机断层扫描(CT)中,金属植入物会引入严重的伪影,导致图像质量降低影响诊断价值。为了对锥束CT中的金属伪影进行校正,提出了一种基于双调和方程的金属伪影校正算法。首先,对含金属伪影的重建图像进行双边滤波和金属阈值分割,获得金属和非金属图像;然后,对二者进行正向投影,获得金属投影区域和先验投影图像;接下来,利用先验投影图像对原始投影进行归一化,并对金属区域进行双调和方程插值修复,获得修复的投影数据;最后,对修复的投影数据去归一化,并利用FDK算法进行重建,再与金属图像融合获得最终的校正图像。为了验证该算法的性能,采用实际拍摄的数据进行金属伪影校正实验。结果表明,与常用的线性插值校正算法和归一化校正算法相比,所提算法ROI区域内图像的均方根误差分别减少了22%和8%,有效地抑制了金属伪影,优于常用的去金属伪影算法。     

非均匀反投影重建方法

本文提出了一种新的计算机断层象(CT)的重建方法--非均匀反投影重建法.该方法 是根据射线上各象元灰度的大小对投影差值进行非均匀加权分配的,而不是采用传统重建方 法中的均匀分配.这种重建方法较好地消除了重建图象中普遍存在的干涉条纹,可明显改善 图象质量.特别是在只有少数几个投影方向数据的非完全投影情况下,该方法效果更好.     

数字X射线图象的成象分析及校正

数字X射线图象的形成是一个X射线信息影象的形成、转换和传递的复杂过程,且通常得到的数字X射线图象的灰度值没有准确的物理意义,在实验的基础上,提出了一种基于朗伯定律的理想图象模型,用以对数字X射线图象进行灰度的校正及线性化,经过这样的处理后,图象的灰度值与吸收体厚度之间的关键近似为线性,这对于人体特征参数的定量分析（如对心脏形态、血流的分析,以及在DSA中对造影剂浓度的分析等）具有重要作用。     

X射线工业CT成像过程复杂伪影抑制方法综述

X射线工业计算机断层(Computerized tomography, CT)技术是一种先进的非接触式无损三维检测技术,能在无损伤情况下以灰度图像的形式对物体内部结构进行全面、详细地分析,在航空航天、工业生产、安检等领域发挥着重要的作用.针对工业CT伪影严重降低图像质量问题,对工业CT成像过程复杂伪影形成机理进行分析,对不同类型伪影抑制方法进行归纳总结.阐述了基于射线衰减、探测器及高密度差异、采样数据及重建等不同过程伪影成因及伪影消除相关算法的最新技术进展,并对近年来人工智能深度学习背景下新兴的基于深度学习及神经网络的工业CT无损检测研究与发展方向进行了总结和展望.     

经MTF修正的骨X射线图象二维谱估计

根据骨X射线图象的谱特征分析来决定骨质疏松的程度,也许可为识别个体的高骨折风险提供一种自动诊断工具,还可作为监控治疗有效性的客观依据。因为各向异性骨小梁在骨X射线图象中表现为弱纹理,其对二维功率谱只贡献很小一部分,从而导致其功率谱被成象系统的调制传递函数和噪声所遮蔽,这里也许可以将发现骨小梁模式的谱特征视为一种反演问题。由于成象系统的调制传递函数和噪声可视为各向同性,并可统一以合成调制传递函数来表达,且各向同性函数在所有的角度上具有相同的投影。因此可在重构骨X射线图象二维功率谱的过程中,利用滤波反投影方法和傅立叶切片定理,并用合成MTF除以投影功率谱,来准确地将骨小梁模式的功率谱提出。在骨X射线图象和合成图象上进行的实验结果表明,由于各向异性的骨小梁,在重构的二维功率谱中有差异显著的峰,因此一旦获得骨小梁的功率谱,就可定位最大峰值频率,以用于估计骨小梁间距和角度,以及其他纹理特征,进而用来描述其各向异性的程度。由于这些纹理特征与骨结构直接相关,因此可用于预测骨折风险。     

基于迭代修正法的CT金属伪影消除

人体内的金属植入物会导致CT重建图像中出现金属伪影。传统的方法利用插值技术替换掉原始投影数据中受金属物污染的数据,以此来达到去除金属伪影的目的,但是容易产生次生伪影。为了解决传统方法的缺陷,提出了一种基于迭代修正的金属伪影消除方法。首先,利用插值修正法获取到一张预处理图像;然后,以该预处理图像为基础进行迭代修正,每一次迭代过程中利用全变分滤波来优化图像质量;经过多次迭代后得到最终图像。实验表明,相比于传统的插值修正法,提出的方法能够更加有效地去除金属伪影,同时较好地抑制次生伪影的产生。     

图像场方向导数的局部区域重建

目的 目前图像修复的研究是以人眼不能察觉为目标,注重视觉效果而不追求重建的准确性。本文提出一种能准确重建图像缺损边缘的重建算法。方法 采用稳定场作为图像局部纹理的数学物理描述,提出基于点源影响函数的图像局部区域重建模型,该模型针对每一个缺损点,计算周围各已知点对它的影响,以期较为准确地重建该点;并根据该场中的方向导数,分析各已知点与缺损点的差异性及相似性,确定一种点源影响函数的计算方法,以实现该重建模型。结果 实验结果表明所提算法与传统修复算法相比,对图像边缘及纹理细节的重建更加清晰,同时保持了较好的整体视觉效果;且重建过程无迭代计算,具有较高的效率。结论 实验结果表明,该算法在重建效率和准确重建方面均取得了较好的成果。     

Performance analysis of structural similarity-based backlight dimming algorithm modulated by controlling allowable local distortion of output image

This paper proposes a global backlight dimming algorithm based on structural similarity that controls the local distortion level of the output image for preserving overall image quality. Unlike existing methods, the proposed method non-iteratively determines a global clipping point that maintains structural similarity for each pixel above the desired level. The proposed method consists of three steps: region analysis, candidate clipping point calculation, and backlight luminance determination. In the first step, the structural similarity is analyzed based on various clipping points. In the second step, the proposed method estimates a candidate clipping point, which preserves the structural similarity index of each pixel above the desired level, based on the estimation of local statistics. Then, the proposed method determines a global clipping point based on a group of candidate clipping points. Finally, the output luminance is determined based on the selected global clipping point. Experimental results showed that the proposed method successfully maintained the structural similarity of each pixel above the desired level. In addition, the proposed method reduced the variation of minimum structural similarity by up to 94.7% compared to the benchmark methods.     

一种基于MRF理论去除Range图象噪声的方法

Ｒａｎｇｅ图象的相关结构可被认为是受噪声干扰的分段光滑面，一旦曲面的参数被确定，那么就可以利用这些参数建曲面，从而大大减少图象中的噪声，Ｒａｎｇｅ图象中，几乎所有象素的统计特性往往与共邻近象的统计特性相关，文中利用ＭａｒｋｏｖＲａｎｄｏｍＦｉｅｌｄ（ＭＲＦ）理论来模型这种相关性，将曲面确定问题转化为一个后验均值求解问题。     

基于像素灰度归类的背景重构算法

背景差法是一种重要的运动检测方法,其难点在于如何进行背景更新.针对该问题,提出一种基于像素灰度归类的背景重构算法,即在假设背景像素灰度以最大概率出现在图像序列的前提下,利用灰度差对相应像素点灰度进行归类,选择频率最高的灰度值作为该点的背景像素值.在背景缓慢变化和突变时,分别利用该算法进行定时和实时背景重构具有时显的优点.仿真结果表明,即使场景中存在运动前景,该算法也能够准确地重构背景,并有效地避免混合现象,从而实现对运动目标的完整提取,以便进一步识别或跟踪.     

Filtering and Reconstruction of LAI Time Series Data by S-G Filter based on Pixel Quality Analysis and Outlier Detection

The Leaf Area Index(LAI) is an important parameter of the canopy structure of the vegetation.There are problems such as missing data and low quality in the MODIS LAI data product because of the influence of atmospheric conditions and other factors,which seriously affects its the application.As a case study of Jiangxi Province,MODIS LAI time series products data set in 2009-2013 were reconstructed by integrating pixel quality analysis,S-G filter and annual sequence abnormal value detection and filtering technology.The results show that the high quality of the broad-leaved forest was the lowest,only 51.76%,and proportion of low quality and retrieval failed pixels reached to 20% to 30%.The integrated filtering method proposed for low quality data set reconstruction.Compared with S-G filtering,the mean value of the reconstructed LAI by the integrated filtering method is more consistent with the original mean for high quality pixel.And its correlation coefficient between the high quality pixel and the original data reaches 0.97.So it has better fidelity.The outliers of the low- and medium-quality pixel reconstruction are filtered,and the area of null value is filled.Meanwhile,the standard deviation is reduced,and the low-value area or abnormal point is better identified and repaired.The overall stability becomes better and it can fit the time series change curve well.     

基于多尺度特征卷积神经网络的目标定位

针对实际应用中诸多数据集标签部分缺失、无定位标注等问题，提出了基于多尺度特征卷积神经网络的弱监督定位算法。其核心思想是利用神经网络分层的特性，在多层卷积层上使用梯度加权类激活映射，生成梯度金字塔模型，并通过均值滤波计算特征质心位置，利用置信强度映射和阈值梯减模块产生连接的像素段，围绕最大边界标注进行弱监督定位。在标准测试集上的实验结果表明，该算法能够在存在大量类别、多尺度图像的情况下完成目标定位，具有较高的精确度。     

一种基于硬件的彩色插值算法

大部分消耗型数字相机都采用单片CCD或CMOS图像传感器,并通过一个彩色滤波阵列来实现,因此在每个像素上只能采集红、绿或蓝一种颜色分量。为获得全彩色图,必须通过插值恢复丢失的色彩信息,通常这个过程被称为demosaicing。鉴于传统方法颜色失真明显,现有方案算法复杂,不利于硬件实现。本文提出一种新的插值算法,引入方向标志位的概念,以及一种平滑的边界检测方案。通过该算法,既可得到满意的插值效果,同时适合数字相机硬件实现,文章最后评估了算法性能。     

Gibbs artifact reduction for POCS super-resolution image reconstruction

The topic of super-resolution image reconstruction has recently received considerable attention among the research community. Super-resolution image reconstruction methods attempt to create a single high-resolution image from a number of low-resolution images (or a video sequence). The method of projections onto convex sets (POCS) for super-resolution image reconstruction attracts many researchers’ attention. In this paper, we propose an improvement to reduce the amount of Gibbs artifacts presenting on the edges of the high-resolution image reconstructed by the POCS method. The proposed method weights the blur PSF centered at an edge pixel with an exponential function, and consequently decreases the coefficients of the PSF in the direction orthogonal to the edge. Experiment results show that the modification reduces effectively the visibility of Gibbs artifacts on edges and improves obviously the quality of the reconstructed high-resolution image.     

Gibbs artifact reduction for POCS super-resolution image reconstruction

The topic of super-resolution image reconstruction has recently received considerable attention among the research community. Super-resolution image reconstruction methods attempt to create a single high-resolution image from a number of low-resolution images (or a video sequence). The method of projections onto convex sets (POCS) for super-resolution image reconstruction attracts many researchers’ attention. In this paper, we propose an improvement to reduce the amount of Gibbs artifacts presenting on the edges of the high-resolution image reconstructed by the POCS method. The proposed method weights the blur PSF centered at an edge pixel with an exponential function, and consequently decreases the coefficients of the PSF in the direction orthogonal to the edge. Experiment results show that the modification reduces effectively the visibility of Gibbs artifacts on edges and improves obviously the quality of the reconstructed high-resolution image.     

Projective reconstruction of ellipses from multiple images

This paper presents a new approach for reconstructing 3D ellipses (including circles) from a sequence of 2D images taken by uncalibrated cameras. Our strategy is to estimate an ellipse in 3D space by reconstructing N(≥5) 3D points (called representative points) on it, where the representative points are reconstructed by minimizing the distances from their projections to the measured 2D ellipses on different images (i.e., 2D reprojection error). This minimization problem is transformed into a sequence of minimization sub-problems that can be readily solved by an algorithm which is guaranteed to converge to a (local) minimum of the 2D reprojection error. Our method can reconstruct multiple 3D ellipses simultaneously from multiple images and it readily handles images with missing and/or partially occluded ellipses. The proposed method is evaluated using both synthetic and real data.