中国数字人数据集三维重建的方法学研究

目的：探索适合于中国数字人数据集特点的图像配准，彩色图像分割以及三维重建方法。方法：首先基于中国数字人连续断层图像的4个定位杆坐标值，对断层图像进行射影变换，消除其射影失真。纠正失真后再将断层图像根据第一个定位杆的质心坐标值裁剪成大小一致的断层图像。其次使用photoshop中knockout滤镜，完成断层图像的分割。最后使用mimics软件完成三维重建。结果：建立了中国数字人女1号髋部骨骼肌精确的三维模型。结论：髋部骨骼肌模型的三维交互可视化可以准确反映该区域复杂的解剖结构及其空间毗邻关系，同时为数字人三维可视化提供了一种精确快速的方法。     

亚像素断层图像配准数据集的建立

目的针对数字化冰冻铣切断层图像的特点,探讨一种实用的高精度图像配准方法,建立基于数字化断层图像的亚像素级配准数据集。方法采用冰冻铣切技术获取成年男性头颈标本的冰冻连续断层图像,在M atlab软件中自动提取定标点图像特征,采用基于2点的刚体变换算法实现图像的自动配准。结果配准后图像定标点与基准配准点的误差小于1个像素,达到亚像素水平。结论采用外定标的图像配准算法可建立亚像素级的配准数据集,定位标记物的准确识别是获得亚像素级配准数据集的保证。     

利用EPID图像金球位置自动跟踪算法研究

摘要：目的：建立一种全新的算法,并利用EPID（Electronic Portal Imaging Device）图像实时读取预埋于体内金球的投影位置,自动重建金球在体内的投影位置坐标。方法：利用所编算法和Matlab软件进行EPID图像读取和处理金球投影位置并自动重建金球在体内的位置坐标。即取通过图像处理和圆曲率识别各自获得的金球投影位置坐标集合的交集为真实的2D坐标,并依此反推到病人体内获得金球位于靶区当中真实的3D重建坐标。结果：本研究共进行了三组实验,每组为两个交角拍摄的EPID片。每组所有EPID片上的金球投影位置探测成功率分别为94．44％,93．75％,94．44％。算法重建的和利用CT扫描在TPS中重建的金球位置的三维坐标最大偏差为2．2mm,最小偏差为0．1mm。该算法于所使用的PC上完成一组（两张交角EPID图像）图像处理和重建的运行时间〈ls。结论：对于多组相交角度拍摄的含金球投影位置的EPID图像,所编算法金球探测成功率比较高,平均为94％,算法重建出的靶区周围金球三维坐标同CT扫描重建出的坐标相比较,偏差在可以接受的范围内。但是该程序参数调节较复杂,而且EPID板接缝处剂量会影响探测成功率．还需要进一步开发自动参数计算程序和消除EPID板接缝处剂量影响。     

构建高分辨率中国数字人男性数据集

目的构建高分辨率中国男性数字人0.2mm断层间距图像数据集。方法取捐献健康男性遗体,经颈动脉红色灌注、低温冷冻定型、专用包埋模具垂直包埋、一次装夹连续等间距铣削、PhaseOneH25数码相机采集连续断层数据以构建数据集。结果共获得9320个RAW格式断层图像数据,分辨率为4080×5440象素,RAW格式单层数据量为27.5M,RAW格式总数据量为260G。结论高分辨率中国数字人男性数据集由于采用了2200万象素的高分辩率数码相机,断层上如血管、脂肪、软骨和骨骼等组织的边界清晰。并可用随机软件对图像的色彩、曝光量以及断层感兴趣区域进行重新处理。     

儿童头部标本薄层断面数据集的建立

目的 以0.1mm层厚建立儿童头部标本薄层断面数据集。 方法 选择中国8岁男孩头颅一个，处于生长发育高峰前期，发育中等，无重大疾病及畸形。标本经过包埋、固定和冷冻30d后，在山东大学医学院断层影像解剖学研究中心实施数控铣切。铣切层厚：0.1mm；铣切环境：-13℃恒低温实验室。标本铣切完毕，使用佳能EOS-1DS MarkII数码相机进行图像采集。 结果 整个头部标本共获得图像2 150张，所有原始图像保存格式为RAW，每张图片约20.4MB，共42.87GB 大小。再分别转换为PSD格式和JPG格式，用于更广泛的数据分享和交流。 结论 成功建立了中国儿童头部断层铣切解剖数据集，该数据集铣切层厚为0.1mm，可以为各种儿童头部相关研究提供解剖学基础。     

双源变螺距螺旋计算机断层成像重建算法

计算机断层成像血管造影术(Computed tomography angiography, CTA)是诊断血管疾病的重要手段.我们提出了一种适用于CTA的双源变螺距螺旋计算机断层成像(Computed tomography, CT)重建算法,投影数据由双源双多层螺旋CT进行采集,采集过程中,螺旋CT的螺距不再是固定的常数,而是随着时间不断地改变.变螺距的双源螺旋CT的重建算法与定螺距的双源双多层螺旋CT重建算法不同之处主要在于采用了新的更一般的轴内插公式.双源变螺距螺旋CT能够更有效地跟踪造影剂,获得具有更高时间分辨率的重建图像.仿真结果表明了该算法的正确性与有效性.     

术中透视图像质量对双平面定位算法精度的影响

目的术中透视图像作为计算机辅助骨科手术(computer assisted orthopedic surgery,CAOS)系统最重要的源头数据,其质量水平对系统定位精度有显著影响。本文定量研究四种典型图像质量参数对双平面定位算法精度的影响。方法使用带有钢珠的标尺作为实验对象,用高精度数字放射平板采集高质量基准图像,通过定量化的人为干预来降低图像质量。使用降低质量后的透视图像作为算法输入,通过双平面定位算法得到靶点位置坐标,与测量得到的真值进行比较,得到定位误差数据。然后使用蒙特卡洛方法分析图像随机误差对定位算法精度的影响。结果相关性分析发现图像分辨率和图像畸变与定位误差有显著相关性,对比度和信噪比对定位误差影响不显著。蒙特卡洛分析发现,[-10,10]像素的随机图像误差,可导致(11.65±9.06)mm的定位误差。结论术中透视图像质量会显著影响空间定位精度,其中分辨率和图像畸变对定位误差有显著影响,对比度和信噪比影响相对较小。因此在保障手术精度的前提下制定合理的术中透视图像质量标准,对于提高CAOS系统定位精度有重要意义。     

"虚拟中国人男性一号"多模态图像配准

目的：解决“虚拟中国人男性一号”CT图像、MRI图像与断层切削图像之间的多模态图像配准问题。材料和方法：根据这三种图像的特点，选择CT图像为基准图像，在对MRI图像进行配准时，通过求解两幅图像梯度特征的最大互信息，搜索出最佳配准参数；在对断层切削图像进行配准时，采用基于解剖结构特征提取的配准方法获取最佳配准参数：最后．根据所得配准参数对待配图进行变换，从而达到配准目的。结果：对头部三种模态图像数据集进行了配准，与高精度手工分割图像数据集进行对比，配准正确率达到95．8％。结论：配准结果准确，解决了“虚拟中国人男性一号”多模态图像配准问题，为数字化虚拟人多模态图像配准提供了参考。     

2D-3D医学图像配准临床数据集标定结果的分析与评价

目的 构建采自临床的2D-3D医学图像配准数据集,是实现各种学习算法应用于实际医疗的重要环节。然而临床数据的获取过程中存在多种不确定因素,致使数据集的标定结果需要分析和评价。本文对采自胸主动脉腔内修复术的一组X线和CT图像的几组标定数据进行分析和评价,并确定正确标定结果。方法 分别采用相似性度量法和投影距离误差法对标定结果进行分析和评价。选用相似性准则,计算CT图像生成的二维数字放射重建图像和X线图像的相似性,相似程度越高,对应的标定值越接近真实值。读取X线图像中的标记物影像位置作为参考位置;将计算得到的CT图像中标记物位置在X线图像上投影,得到投影位置;计算参考位置和投影位置的距离,距离值越小,对应的标定值越接近真实值。结果 提供的几组标定数据,在比较数字放射重建图像和X线图像相似性方面,相似度接近,没有明显指向性;而投影距离误差法的分析结果指向性明显,能够定量描述标定结果的优劣。主要原因在于各组标定值之间差别不突出;生成的数字放射重建图像和X线图像之间模态差异较大等。结论 投影距离误差法是评价2D-3D医学图像配准数据集标定结果的有效手段。另外,若提供的标定结果计算数据差异明显,或...     

基于三维模型的断层图像结构定位的初步研究

目的:研究利用初步建立的三维模型确定二维断层图像感兴趣结构像素坐标的方法.方法:通过Photoshop图像处理软件绘制断面图像,使用可视化工具包VTK的移动立方体表面重建算法,在VC++6.0的编译环境下对其进行三维重建以及立体显示.用自行开发的坐标转换处理程序对三维模型上提取的坐标值进行处理,计算断层图像相应结构的像素坐标.结果:建立了一个表面带有S型凹槽的三维模型,通过计算三维模型上凹槽结构的一系列坐标,得到二维断层图像上相应结构的像素坐标点.结论:本研究以VTK重建的三维模型为基础,提出了一种利用已建成的三维模型来指导二维断层图像结构定位的方法,为人体复杂结构的分割与修正以及某些在二维断层图像上无法识别的结构的定位提供了一种新的手段.     

A faster method for 3D/2D medical image registration--a simulation study

3D/2D patient-to-computed-tomography (CT) registration is a method to determine a transformation that maps two coordinate systems by comparing a projection image rendered from CT to a real projection image. Iterative variation of the CT's position between rendering steps finally leads to exact registration. Applications include exact patient positioning in radiation therapy, calibration of surgical robots, and pose estimation in computer-aided surgery. One of the problems associated with 3D/2D registration is the fact that finding a registration includes solving a minimization problem in six degrees of freedom (dof) in motion. This results in considerable time requirements since for each iteration step at least one volume rendering has to be computed. We show that by choosing an appropriate world coordinate system and by applying a 2D/2D registration method in each iteration step, the number of iterations can be grossly reduced from n6 to n5. Here, n is the number of discrete variations around a given coordinate. Depending on the configuration of the optimization algorithm, this reduces the total number of iterations necessary to at least 1/3 of it's original value. The method was implemented and extensively tested on simulated x-ray images of a tibia, a pelvis and a skull base. When using one projective image and a discrete full parameter space search for solving the optimization problem, average accuracy was found to be 1.0 +/- 0.6(degrees) and 4.1 +/- 1.9 (mm) for a registration in six parameters, and 1.0 +/- 0.7(degrees) and 4.2 +/- 1.6 (mm) when using the 5 + 1 dof method described in this paper. Time requirements were reduced by a factor 3.1. We conclude that this hardware-independent optimization of 3D/2D registration is a step towards increasing the acceptance of this promising method for a wide number of clinical applications.     

Relationship Between the Contrast Effects of Raw Data Projection Images from Three-Dimensional Digital Subtraction Angiography and the Optimal Volume Rendering Parameters

Volume rendering (VR) is a technique commonly used for the reconstruction of three-dimensional (3D) digital subtraction angiography (DSA) images, and the rendering parameters greatly affect the characteristics of the 3D image. This study aimed to test whether the optimal VR parameters for 3D DSA could be estimated from the contrast effects in rotational two-dimensional (2D) DSA images acquired using 3D DSA. Simulated blood vessels filled with various concentrations of contrast medium were scanned, and the 3D DSA data sets were reconstructed. The syngo AX vessel analysis software that was able to analyze 3D DSA VR image was used for objective measures. Raw data projection images of the 3D DSA data sets in which the mean diameter was calculated as a true value by the software at nine different thresholds for vessel segmentation were selected. In each image set, five images of all 133 rotational 2D DSA images were selected, and the contrast-enhanced area was extracted using a region-growing algorithm. Mean values and standard deviations of each contrast-enhanced area were calculated, and as the thresholds for vessel segmentation of the software increased by 500 every time, significant differences were observed in the mean values (P < 0.01). This optimal threshold can be applied to the window settings of the VR technique. Therefore, the optimal VR parameters for 3D DSA may be determined by analyzing the contrast effects of the raw data projection images, and user-dependent over- and underestimations of 3D DSA VR images also may be prevented.     

基于傅里叶梅林变换的图像配准方法

目的:提出一种新的配准框架用于图像引导放射治疗系统中的2D/3D图像配准,有效降低传统方法迭代搜索时间,同时保证放射治疗要求的配准精度。方法:利用傅里叶梅林变换方法对正侧位kV图像与对应方位参考CT图像生成的数字重建放射影像（DRR）进行粗配准,根据傅里叶梅林变换计算得到的二维平移向量以及放射治疗系统的机械几何参数反推出参考CT图像的三维空间位置偏差,更新正侧位的DRR图像,最后通过正侧位kV图像与DRR图像的相似度进行精配准达到临床需求。结果:采用临床金标准数据验证方法的配准性能,实验结果表明,配准误差为0.576 5 mm,平均运行时间为3.34 s。结论:该方法鲁棒性强,对图像的噪声不敏感,人工干预少,可满足临床应用的需求。     

由CT图像和彩色数码照片合成真实感人脸

为了预测整复外科手术后的效果,我们提出了一种从3张彩色数码照片和由CT图像重建得到的3D人脸模型合成高清晰度真实感人脸的方法.通过手工选取的2D和3D特征点之间的对应关系,采用了遗传算法估计了从3D人脸到2D照片的投射参数,从而实现了3D人脸到2D照片的配准.采用多重纹理映射的方法对纹理进行融合,在消除人脸侧面拼接裂缝的同时获得了高清晰的人脸纹理.     

Automated detection of microcalcification clusters for digital breast tomosynthesis using projection data only: a preliminary study

Digital breast tomosynthesis (DBT) is a promising modality for breast imaging in which an anisotropic volume image of the breast is obtained. We present an algorithm for computerized detection of microcalcification clusters (MCCs) for DBT. This algorithm operates on the projection views only. Therefore it does not depend on reconstruction, and is computationally efficient. The algorithm was developed using a database of 30 image sets with microcalcifications, and a control group of 30 image sets without visible findings. The patient data were acquired on the first DBT prototype at Massachusetts General Hospital. Algorithm sensitivity was estimated to be 0.86 at 1.3 false positive clusters, which is below that of current MCC detection algorithms for full-field digital mammography. Because of the small number of patient cases, algorithm parameters were not optimized and one linear classifier was used. An actual limitation of our approach may be that the signal-to-noise ratio in the projection images is too low for microcalcification detection. Furthermore, the database consisted of predominantly small MCC. This may be related to the image quality obtained with this first prototype.     

血管造影图像序列中冠状动脉的三维运动估计

提出了由两个角度的单面血管造影图像序列估计冠状动脉骨架树三维运动的算法。首先对冠状动脉造影图像序列进行二维预处理和二维运动估计。然后根据冠脉造影系统的透视投影模型得到两幅不同角度的造影图像之间的几何变换关系，以及空间点三维坐标的计算方法。最后，在对整个图像序列进行分析的过程中，将三维运动估计与重建结合起来，得到各骨架点的三维运动向量。采用临床得到的冠状动脉造影图像序列对算法进行了验证，并分析了误差源。     

Digital image analysis of pan-cytokeratin stained tumor slides for evaluation of tumor budding in pT1/pT2 colorectal cancer: Results of a feasibility study

Tumor budding is an independent prognostic factor in colorectal cancer. However, varying degrees of interobserver agreement and reproducibility challenges the use of tumor budding in diagnostics. Immunohistochemical staining of tumor slides with pan-cytokeratin visualizes the budding tumor cells and has been suggested to improve reproducibility. Here we demonstrate the methodology of tumor budding assessment using digital image analysis based on tumor slides stained for pan-cytokeratin, and investigate interobserver agreement, agreement between manual and digital assessment methods and digital reproducibility between users.Tumor slides from 126 patients with pT1/pT2 colorectal cancer were stained with pan-cytokeratin and tumor budding at the invasive tumor front was assessed by conventional manual microscopy. A digital image analysis algorithm for identification and quantification of budding tumor cells was developed and tested on the pan-cytokeratin stained slides.Manual assessment of tumor budding using pan-cytokeratin stained tumor slides exhibited high correlations (Spearman Rank 0.84-0.89, p?<?0.001),excellent agreement between observers (Intra-class correlation coefficient (ICC): 0.86 -0.87) and 2.20 higher odds for regional metastases with increasing budding counts (p?=?0.017). Digital image analysis correlated well to manual assessment (Spearman Rank 0.71-0.88) and agreement between the two methods was good (ICC 0.62-0.82). However, only a trend towards increased odds for metastatic progression was found for the adjusted digital estimates (p?=?0.076). Digital estimates were higher than manual estimates, demonstrated by a systematic median difference of 3–4.5 buds. Image analysis was highly reproducible between users of the algorithm (ICC 0.98). In conclusion, assessment of tumor budding using pan-cytokeratin stained tumor slides is a method with high correlation and agreement between observers. Digital image analysis quantifies budding tumor cells in high agreement with manual estimates, but approval of the digital slides by a pathologist is mandatory. The method qualifies for further investigation.     

Image intensifier distortion correction

A method is described for determining and correcting the spatial distortion affecting images acquired with image-intensified video systems. The distortion is separated into two physically distinct components, a predominant one originating from the projection of the x-ray image onto the curved input phosphor, and a smaller component corresponding to the mapping from the input phosphor to the output phosphor and the digital image matrix. The former is determined geometrically and the latter is modeled with four empirical parameters. A method for determining the two components from calibration images of a grid phantom is presented. The model of the image intensifier distortion was found to be in good agreement with the measured distortion in the images. The correction is applicable to arbitrary angulations of the x-ray beam with respect to the image intensifier tube, and consequently is directly applicable to digital tomosynthesis. The correction allows one to obtain accurate positional information with fluorography and may also be useful in radiation therapy treatment planning and quantitative digital subtraction angiography.     

Standardization for image characteristics in telemammography using genetic and nonlinear algorithms

As the soft copy reading and computer assisted diagnosis (CAD) in mammography become more and more important, the standardization of digital images becomes paramount. Telemammography and telemedicine requires the standardization for image characteristics, such as image resolution, bit-depth and intensity response. Soft copy reading and CAD in mammography are both dependent on the characteristics of the source of the digital data, either direct digital mammography or digitized screen-film mammography. An algorithm developed on images from one database may not perform well as on images from another database (with a different digitization). In this paper, we describe two methods based on a genetic algorithm and a nonlinear algorithm for standardization of digitized and digital mammography. The proposed standardization techniques are based on geometric and intensity transformations that are discovered using a set of calibration images. A set of transformation algorithm is used to search for the best standardization.     

Human efficiency for symmetry detection on 3D bumpy surfaces

Based on the ideal observer analysis, we investigated sampling properties of image information used by human visual system, for symmetrical pattern discrimination on 3D bumpy surface. There were three models of ideal observer (IO) to perform the task: 2D-IO using 2D projection image (i.e., retinal image), 2.5D-IO using image transformed to canonical view, and 3D-IO using recovered pattern image of 2D plane. We measured discrimination thresholds on the task for each IO model and subjects, and calculated human statistical efficiency relative to each ideal observer. The results indicated for the detection of a diagonal symmetry in the bumpy surface that human performance was similar to 3D-IO. This implies that human observers use the structure of the bumpy surface to detect the diagonal symmetry.