改进的三维可视化用光线投射算法

把图像处理、光线投射与包围体技术有机结合，提出了一种提高成像质量和速度的三维可视化新方法。该方法利用物体空间的包围体算法来减少追踪光线的数量，加快了绘制速度。通过实际的医学胸部CT图像的三维重建实验，取得了较好的三维显示效果和速度，验证了改进的光线投射算法对胸部CT图像的快速三维可视化问题的有效性。     

组织切片图像的可视化技术及应用

目的:研究和开发序列组织切片图像的三维显示技术。方法:首先利用序列切片中上下层切片相关点的对应关系,用最小二乘方法实现上下层切片图像的几何变换配准,进而用序列图像插值算法完成组织切片图像的体数据重建。结果:在对重点显示目标分类的基础上,用基于物空间SHEAR-WARP快速直接体视方法对体数据进行三维显示,组成一套医学图像可视化系统。结论:该系统能对胆管癌等序列切片图像进行三维重建显示。     

基于工业CT断层图像的三维可视化

工业CT图像三维可视化能够对工业构件提供真实、直观的反映。体绘制技术可以显示工业CT三维数据的整体特征和内部细节信息。根据光线投射算法的特点，采用对原始数据场进行最大熵原则的预处理的方法，加快了绘制速度，在一定程度上改进了光线投射算法，取得了较好的显示效果。     

基于像素采样不足的伪迹处理方法

在实际的三维锥束CT重建中，受扫描结构的限制，当连续X射线穿过物体时，像素对射线的采样间隔不能满足Nyqust采样定理，这样通道信号产生混叠，使得重建图像出现伪迹。本论文针对这一问题，在用splatting算法计算代数重建算法权系数的基础上，分析了伪迹产生的原因并对对应的内插函数作了适当的变换，提出了适用于3-D锥束圆轨迹CT扫描系统的去伪处理方法。结果表明，经过去伪处理后，重建图像的分辨率明显提高。     

直线轨迹扫描断层成像中的图像重建与恢复方法

基于直线轨迹扫描的X射线断层成像方式,具有成像速度快、无物体重叠问题且造价较低等优点,在快速行李物品安检、大物体检查等领域具有较大应用潜力。不完备数据(有限角度)的图像重建及恢复是该成像方式的关键难题。本文介绍了一种适用于直线轨迹扫描断层成像的图像重建及恢复方法:在直接解析重建的基础上,利用linogram技术和总变分(Total Variation)正则化准则,实现快速稳定的外插迭代以补偿缺失数据,从而在保持图像边缘的同时,提高了图像像素值准确性,减小了图像伪影。     

工业过程断层图像的三维动态可视化

提出了一种针对工业过程断层图像的三维动态可视化方法,可用于对工业过程的辅助监控。该方法采用基于光线投射的体可视化技术,已在MITK(Medical Imaging ToolKit,一个用于医学影像处理与分析的C++类库)中实现。该方法使用不同的颜色和阻光度系数来区分反应容器或管道中的不同物质,从而为容器或管道中不同物质的混合反应过程提供一个动态的显示。实验结果证明该方法是可行的,并且其性能也是可接受的,若再辅之以体绘制算法的硬件加速,该方法可用于实时的工业断层成像系统中。     

光学投影断层成像在离体样本中的应用研究

目的利用光学投影断层成像系统对小鼠离体心脏和肝脏进行三维成像,进而通过考察三维成像效果研究光学投影断层成像的应用价值。方法对小鼠进行心脏灌注,分别取出心脏、肝脏组织,并对组织进行琼脂包埋、脱水和透明化处理,得到心脏和肝脏样本。利用光学投影断层成像系统采集心脏和肝脏的透射数据,并通过滤波反投影的方法得到心脏和肝脏的三维图像。分析得到的三维图像,研究光学投影断层成像的应用价值。结果通过上述实验方法得到心脏和肝脏样本,利用光学投影断层成像系统得到具有较高清晰度和空间分辨率的样本三维结构图像。结论利用光学投影断层成像技术可获得具有高清晰度的生物组织器官的三维结构图像,在组织分辨率水平上获得小动物器官的数据信息,为尺度在1~10 mm的生物样本提供了有力的研究手段,必将大大推动生物科学基础研究的发展。     

基于双边滤波的POCS超分辨率图像序列重建算法

针对传统凸集投影(POCS)超分辨率图像序列重建算法使用高斯滤波器来估计点扩散函数(PSF)导致边缘模糊的现象,本文采用双边滤波器来估计PSF,双边滤波方法是将高斯函数和图像亮度信息相乘,优化后的权系数再与图像作卷积,使重建后的图像边缘得到很好保持。运动估计是POCS超分辨率图像序列重建算法的关键技术,本文选择结合图像金字塔的光流估计对图像序列进行配准,得到更加精确的估计结果。实验表明,可以使重建图像取得良好的视觉效果。     

基于图像自适应融合的超分辨率重建算法

本文提出了一种基于图像块自适应融合的序列图像超分辨率重建算法。算法使用低分辨率序列图像中的互补信息重建生成高分辨率图像。为了保证重建初始估计尽可能接近真实场景,配准后的序列图像按照图像块梯度信息自适应的融合生成高分辨率初始估计图像。算法采用误差反向投影的方法对高分辨率图像迭代校正,生成超分辨率重建最终结果。实验证明,本文提出的超分辨率重建算法能够在增加图像细节信息的同时重建出更加自然真实的高分辨率图像。     

DART-TV：高精度离散断层图像重建算法

本文提出了一种基于离散代数迭代（DART）和全变差（TV）最小化的高精度图像重建算法,即DART．TV。该算法适用于工业CT辅助制造中对于单一材料物体的尺寸检测,特别是在物体对某些角度无法被穿透的情况。传统的CT重建算法如FBP或ART无法有效处理这类数据缺失的病态问题,重建图像分辨率差。本文提出的算法的核心思想是：（1）利用DART的结果限制重建物体的支撑域,改善问题的病态性;（2）迭代过程中施加TV约束,以抑制噪声;（3）每次迭代后自动更新支撑域,不断逼近真实解,使得图像边缘更加清晰。初步的模拟结果和实验结果显示,DART-TV算法在有限角情况下可以得到更精确的重建图像边缘,并且该算法对于投影噪声和灰度值估计误差有更好的鲁棒性,可以应用于实际工程。     

用于颅颌面种植外科的三维立体可视化系统

目的论文对在个人PC机上对颅颌面医学图像的可视化进行研究,开发出用于颅颌面种植外科的CT断层图像三维可视化系统。方法应用3D纹理映射的方法对颅颌面CT断层图像进行三维重建,并得到其XYZ方向上的切面图像。然后利用边缘检测的方法得到眼眶的位置,并在眼眶位置进行扇形切割,得到18幅扇形切割图像。结果该系统可以帮助医生进行手术前的骨质、骨量分析,设计种植手术的过程及模拟、预测手术的种植效果。结论建立的颅颌面种植外科三维可视化系统以其硬件配置合理、软件设计新颖、多维视角、便捷快速精确等为特色。具有较强的应用价值。     

螺旋CT三维重建技术对非小细胞肺癌放射治疗靶区勾画的研究

目的 研究螺旋CT三维重建技术对非小细胞肺癌放射治疗靶区勾画准确性的影响。方法 应用螺旋CT扫描,图像后处理应用TOMOCON图像融合系统。对肺内病灶分别以常规二维CT及二维CT结合螺旋CT三维重建技术进行肿瘤靶区勾画、体积计算及靶区横径及纵径比较。结果 48例肺癌病人CT定位扫描后常规二维CT靶区勾画结果:靶区体积(GTV)18.85~380.63 cm³,中位体积103.60 cm³;二维CT结合螺旋CT三维重建勾画结果:靶区体积(GTV)24.05~420.57 cm³,中位体积130.55 cm³。两组数据经配对t检验,分析结果P=0.037,有统计学意义。两组勾画结果显示横径平均差异为(1.5±0.5)mm。纵径平均差异为(3.0±1.5)mm。结论 螺旋CT三维重建技术在肺癌放射治疗靶区勾画中的应用价值是值得肯定的。可以作为常规CT横断面扫描图像勾画的一个有力补充,使得肺癌放射治疗靶区勾画的准确性大大提高。     

Volume Reconstruction Techniques Improve the Correlation Between Histological and in vivo Tumor Volume Measurements in Mouse Models of Human Gliomas

Assessment of therapy efficacy using animal models of tumorigenic cancer requires the ability to accurately measure changes in tumor volume over the duration of disease course. In order to be meaningful, in vivo tumor volume measurements by non-invasive techniques must correlate with tumor volume measurements from endpoint histological analysis. Tumor volume is frequently assessed by endpoint histological analyses approximating the tumor volume with geometric primitives such as spheroids and ellipsoids. In this study we investigated alternative techniques for quantifying histological volume measurements of tumors in a xenograft orthotopic mouse model of human glioblastoma multiforme, and compared these to in vivo tumor volume measurements based on magnetic resonance imaging (MRI) data. Two techniques leveraging three-dimensional (3D) image analysis methods were investigated. The first technique involves the reconstruction of a smoothed polygonal model representing the tumor volume from histological section images and is intended for accuracy and qualitative assessment of tumor burden by visualization, while a second technique which approximates the tumor volume as a series of slabs is presented as an abbreviated process intended to produce quantitatively similar volume measurements with a minimum of effort required on behalf of the investigator. New software (QuickVol) designed for use in the first technique, is also discussed. In cases where tumor growth is asymmetric and invasive, we found that 3D analysis techniques using histological section images produced volume measurements more consistent with in vivo volume measurements based on MRI data, than approximation of tumor volume using geometric primitives. Visualizations of the volumes represented by each of these techniques qualitatively support this finding, and suggest that future research using mouse models of glioblastoma multiforme (genetically engineered or xenograft) will benefit from the use of these or similar alternative tumor volume measurement techniques.     

CT扫描/重建参数对三维治疗计划系统影像的影响

[目的]研究CT模拟定位中,CT扫描/重建参数对三维治疗计划重建的三维假体的几何精确度的影响.[方法]在西门子CT模拟机（Somatom plus 4）上扫描自制模体,扫描所得图像登记到ADAC三维治疗计划系统重建成三维假体,测量假体的相关坐标数据并与模体的实际数据相比较;对Catphan 412模体扫描并测量各组图像的实际层厚,讨论实际层厚的变化对计划系统中登记影像的几何精度影响.[结果]（1）CT扫描所采用的不同扫描/重建参数对三维计划系统中重建的三维假体的左右及上下方向的几何精度影响不大,但对重建假体的前后方向（即模体扫描的步进方向）的几何精度有一定的影响.（2）CT扫描所采用的螺距及重建模式会对层厚敏感度曲线（SSP）半高宽值产生影响,该变化对重建假体的前后方向几何精度同样有一定的影响.[结论]重建CT图像的前后方向的几何误差是随着扫描层厚增加而增加,主要是由于CT扫描的部分容积效应影响.单纯增加螺距或使用360度线性内插（Wide）重建模式,都会引起CT图像实际层厚的增加,引起更大的容积效应影响.同时部分容积效应也会导致三维治疗计划系统中数字重建影像（DRR）分辨率的降低.     

系列截面法层间距对晶粒体积和表面积测量的影响规律

系列截面分析方法是显微组织3D定量表征及3D重建的重要方法,为兼顾结果的准确性和效率,系列截面法中层间距的选取非常重要。本文选取不同仿真尺度(300×300×300和400×400×400)的Potts模型Monte Carlo仿真晶粒组织为研究对象,研究了系列截面层间距对三维个体晶粒的体积和表面积测量结果的影响规律,结果表明:系列截面的间距值h<0.111珋l和h<0.135珋l(珋l为晶粒平均截线长度)时,分别可保证晶粒的体积和表面积测量的相对误差小于5%和10%。     

A new, accurate and conventional five-point method for quantitative evaluation of ascites using plain computed tomography in cancer patients

BACKGROUND: To assess the exact response of the malignant ascites to the treatment, the objective measurement of the volume of ascites is essential. We have developed a simple method to measure the volume of ascites by using standard abdomino-pelvic computed tomography (CT). The aim of the study is to validate the accuracy of the measurements by comparing them with the standard volume calculation by using 3D-CT. METHODS: Twelve consecutive patients with cancer who had measurable ascites underwent 15 helical CT examinations. On conventional CT images, the thickness of ascites in centimeters was measured in three planes such as the bilateral subphrenic space (A and B), the bilateral paracolic space (C and D) and the pre-bladder space (E), and the average thickness: (A + B + C + D + E)/5 was then multiplied by the area of standard abdominal cavity in the anterior projection, that was assumed to be 1000 cm(2), to yield the volume of ascites: (A + B + C + D + E) x 200 (ml). The volume of ascites was compared with the exact volume, that was obtained from 3D-CT with the volume rendering method. RESULTS: The volume of ascites measured by the present method and the volume rendering method ranged from 140 to 4040 ml and from 86 to 4279 ml, respectively. The correlation was statistically significant with a correlation coefficient of 0.956 (P < 0.01) using the Spearman's rank correlation. In 13 examinations with the exact volume > or =300 ml, the average ratio of the absolute difference in the volume was 12.9 +/- 13.9% as compared with 62.8 and 162.0% in two examinations with the exact volume < 300 ml. CONCLUSION: The preliminary study indicated that the present five-point method using a conventional CT was accurate in patients with the volume of ascites > or =300 ml. Because this procedure is simple and easy to perform, it should be feasible in many hospitals for the follow-up of ascites after treatment.     

基于3D Slicer软件的蝶鞍区三维重建技术在经蝶内镜垂体瘤切除术中的应用价值

目的:探讨基于3D Slicer软件的蝶鞍区三维重建对经蝶内镜垂体瘤切除术的指导作用。方法:将2017年10月至2019年3月徐州医科大学附属医院神经外科收治的70例垂体瘤病人随机分为两组:术前重建组和传统手术组。术前重建组术前使用图像处理软件3D Slicer对影像学数据进行蝶鞍区骨性标志、垂体瘤瘤体及颈内动脉、视神经结构的三维重建,并进行术前规划。将三维重建所显示的鞍区解剖标志与术中内镜下所见术野图像进行对比分析,评价三维重建图像与术野图像的吻合度。传统手术组不进行三维重建,按常规流程实施手术。比较术前重建组与传统手术组在手术时间、术中出血量、术后住院时间及术后并发症发生率方面的差异。结果:三维重建与术中内镜对上鼻甲、中鼻甲、鼻中隔完全显示率均为100%。三维重建对于蝶窦开口、视神经管隆起、颈内动脉隆起、颈内动脉-视神经隐窝的完全显示率分别为100%、88%、94%、94%,均显著高于术中内镜所见（P＜0.05）。术前重建组在手术时间、术中出血量、术后住院时间、术中脑脊液漏发生率方面优于传统手术组（P＜0.05）,而视神经损伤、颈内动脉损伤、术后脑脊液漏发生率、术后尿崩的发生率上,两组没有明显差异（P＞0.05）。结论:基于3D Slicer的三维重建技术对鞍底解剖标志的显示具有较高的还原性和准确度,有助于提前预判手术场景,开展手术教学。同时对于未配备术中导航的医疗机构,可以一定程度上提高手术精确度和安全性,降低并发症发生率。     

组织切片图像的Disector计算机自动计数及三维显示

本文介绍了自行开发的解决Disector应用问题的计算机图像处理软件系统及关键技术、算法。本系统能依照Disector原理对多层组织细胞的电镜及光镜图像中的目标进行正确的分割 ,各层之间图像的人工交互及自动配准。针对每个目标的上下层关系实现目标的Disector自动计数及参数计算 ,并对每个目标进行体数据的提取和三维重建显示。利用该系统可在三维空间对粒子进行无偏差计数 ,同时极大地减少复杂的手工操作。文中同时还给出了用该系统对小白鼠神经细胞个数的Disector分析 ,肾小球个数统计分析和三维重建显示实例。     

Use of three-dimensional radiation therapy planning tools and intraoperative ultrasound to evaluate high dose rate prostate brachytherapy implants

PURPOSE: We performed a pilot study to evaluate the quality of high dose rate (HDR) prostate implants using a new technique combining intraoperative real-time ultrasound images with a commercially available 3-dimensional radiation therapy planning (3D RTP) system. METHODS AND MATERIALS: Twenty HDR prostate implants performed by four different physicians on a phase I/II protocol were evaluated retrospectively. Radiation therapy (RT) consisted of pelvic external beam RT (EBRT) to a dose of 46 Gy in 2-Gy fractions over 5 weeks and 2 HDR implants (prescribed dose of 950 cGy per implant). Our in-house real-time geometric optimization technique was used in all patients. Each HDR treatment was delivered without moving the patient. Ultrasound image sets were acquired immediately after needle placement and just prior to HDR treatment. The ultrasound image sets, needle and source positions and dwell times were imported into a commercial computerized tomography (CT) based 3D RTP system. Prostate contours were outlined manually caudad to cephalad. Dose-volume histograms (DVHs) of the prostate were evaluated for each implant. RESULTS: Four patients with stage T2a carcinoma, 4 with stage T2b, and 3 with stage T1c were studied. The median number of needles used per implant was 16 (range 14-18). The median treated volume of the implant (volume of tissue covered by the 100% isodose surface) was 82.6 cc (range 52.6-96.3 cc). The median target volume based on the contours entered in the 3D RTP system was 44.83 cc (range 28.5-67.45 cc). The calculated minimum dose to the target volume was 70% of the prescribed dose (range 45-97%). On average 92% of the target volume received the prescribed dose (range 75-99 %). The mean homogeneity index (fraction of the target volume receiving between 1.0 to 1.5 times the prescribed dose) was 80% or 0.8 (range 0.55-0.9). These results compare favorably to recent studies of permanent implants which report a minimum target volume dose of 43% (range 29-50%) and an average of 85% of the target volume (range 76-92%) receiving the prescribed dose. CONCLUSIONS: The feasibility of evaluating HDR prostate implants using ultrasound images (acquired immediately prior to treatment) with a commercially available 3D RTP system was established. The dosimetric characteristics of these HDR implants appear to be substantially different compared to permanent implants. These developments allow quantitative evaluation of the dosimetric quality of HDR prostate treatments. Future studies will examine any correlation between the dosimetric quality of the implant and clinical/biochemical outcomes.