基于三维全卷积网络的肝脏和肝癌分割算法研究

为了解决计算机断层扫描(computed tomography,CT)影像中肝脏和肝癌的准确分割问题,提出了基于三维全卷积网络的肝脏分割算法和肝癌分割算法。肝脏分割算法和肝癌分割算法都采用Vnet网络进行分割。在肝脏分割算法中,采用了形态学方法进行后处理,提高了肝脏分割准确率。在肝癌分割算法中,采用了组合损失函数训练Vnet网络,使得Vnet网络更好地收敛,并加入后处理提高了肝癌分割准确率。为了验证算法的性能,采用MICCAI 2017 Liver Tumor Segmentation Challenge(LiTS)数据集进行了肝脏分割和肝癌分割的5折交叉验证实验。肝脏分割算法在测试集的平均分割准确率为0.9510,高于Unet网络和3D Unet网络;肝癌分割算法的平均分割准确率为0.712。实验结果表明,肝脏分割算法可以准确地对肝脏进行分割,肝癌分割算法也达到了较高的准确率。     

Automated PET-guided liver segmentation from low-contrast CT volumes using probabilistic atlas

The use of the functional PET information from PET-CT scans to improve liver segmentation from low-contrast CT data is yet to be fully explored. In this paper, we fully utilize PET information to tackle challenging liver segmentation issues including (1) the separation and removal of the surrounding muscles from liver region of interest (ROI), (2) better localization and mapping of the probabilistic atlas onto the low-contrast CT for a more accurate tissue classification, and (3) an improved initial estimation of the liver ROI to speed up the convergence of the expectation-maximization (EM) algorithm for the Gaussian distribution mixture model under the guidance of a probabilistic atlas. The primary liver extraction from the PET volume provides a simple mechanism to avoid the complicated pre-processing of feature extraction as used in the existing liver CT segmentation methods. It is able to guide the probabilistic atlas to better conform to the CT liver region and hence helps to overcome the challenge posed by liver shape variability. Our proposed method was evaluated against manual segmentation by experienced radiologists. Experimental results on 35 clinical PET-CT studies demonstrated that our method is accurate and robust in automated normal liver segmentation.     

支持向量机分类的交互式肝脏肿瘤分割方法

提出了一种交互式分割传统CT图像肝脏肿瘤的方法。首先对CT切片进行预处理,包括肝脏薄壁组织分割及其对比增强处理,通过分水岭转换后肝脏体积被分成许多集水盆地。然后,在用户选择种子点上训练支持向量机分类来抽取肝脏肿瘤,而在分水岭转换后产生的每个小区域基础上,计算对应的用于训练和预测的特征向量。最后,在整个分割二级制体数据中执行一些形态学操作,重新定义支持向量机分类的粗糙分割结果。实验结果表明：改进方法提高了诊疗的准确性、有效性,以及在临床应用中的可行性。     

基于改进窄带水平集的三维肝脏肿瘤分割

三维肝脏肿瘤识别是当前研究的热点问题,如何准确快速地从腹部CT序列中分割出肝脏肿瘤是肝部病变诊断的基础。针对水平集方法在进行分割时收敛速度较慢,设置窄带宽度固定不灵活的缺点,先利用分水岭算法,对肝脏图像进行“过分割”,搜索初始轮廓所在的分水岭块作为窄带区域进行标记,在窄带区域内用水平集算法使初始轮廓线收敛至准确轮廓。再以其边缘作为相邻CT序列的肿瘤初始轮廓,找出初始轮廓线所在的分水岭块,构成新的窄带,用水平集算法对轮廓线进行迭代分割出肿瘤。重复该过程,直至完成整个肝脏序列图像的肿瘤图像分割,进行三维重建。     

Kernel sparse representation for MRI image analysis in automatic brain tumor segmentation

The segmentation of brain tumor plays an important role in diagnosis, treatment planning, and surgical simulation. The precise segmentation of brain tumor can help clinicians obtain its location, size, and shape information. We propose a fully automatic brain tumor segmentation method based on kernel sparse coding. It is validated with 3D multiple-modality magnetic resonance imaging (MRI). In this method, MRI images are pre-processed first to reduce the noise, and then kernel dictionary learning is used to extract the nonlinear features to construct five adaptive dictionaries for healthy tissues, necrosis, edema, non-enhancing tumor, and enhancing tumor tissues. Sparse coding is performed on the feature vectors extracted from the original MRI images, which are a patch of m×m×m around the voxel. A kernel-clustering algorithm based on dictionary learning is developed to code the voxels. In the end, morphological filtering is used to fill in the area among multiple connected components to improve the segmentation quality. To assess the segmentation performance, the segmentation results are uploaded to the online evaluation system where the evaluation metrics dice score, positive predictive value (PPV), sensitivity, and kappa are used. The results demonstrate that the proposed method has good performance on the complete tumor region (dice: 0.83; PPV: 0.84; sensitivity: 0.82), while slightly worse performance on the tumor core (dice: 0.69; PPV: 0.76; sensitivity: 0.80) and enhancing tumor (dice: 0.58; PPV: 0.60; sensitivity: 0.65). It is competitive to the other groups in the brain tumor segmentation challenge. Therefore, it is a potential method in differentiation of healthy and pathological tissues.     

肝脏肿瘤CT图像深度学习分割方法综述

肝脏肿瘤的精确分割是肝脏疾病诊断、手术计划和术后评估的重要步骤。计算机断层成像（computed tomography,CT）能够为肝脏肿瘤的诊断和治疗提供更为全面的信息,分担了医生繁重的阅片工作,更好地提高诊断的准确性。但是由于肝脏肿瘤的类型多样复杂,使得分割成为计算机辅助诊断的重难点问题。肝脏肿瘤CT图像的深度学习分割方法较传统的分割方法取得了明显的性能提升,并获得快速的发展。通过综述肝脏肿瘤图像分割领域的相关文献,本文介绍了肝脏肿瘤分割的常用数据库,总结了肝脏肿瘤CT图像的深度学习分割方法：全卷积网络（fully convolutional network,FCN）、U-Net网络和生成对抗网络（generative adversarial network,GAN）方法,重点给出了各类方法的基本思想、网络架构形式、改进方案以及优缺点等,并对这些方法在典型数据集上的性能表现进行了比较。最后,对肝脏肿瘤深度学习分割方法的未来研究趋势进行了展望。     

Fuzzy Markovian Segmentation in Application of Magnetic Resonance Images

In this paper, we present a fuzzy Markovian method for brain tissue segmentation from magnetic resonance images. Generally, there are three main brain tissues in a brain dataset: gray matter, white matter, and cerebrospinal fluid. However, due to the limited resolution of the acquisition system, many voxels may be composed of multiple tissue types (partial volume effects). The proposed method aims at calculating a fuzzy membership in each voxel to indicate the partial volume degree, which is statistically modeled. Since our method is unsupervised, it first estimates the parameters of the fuzzy Markovian random field model using a stochastic gradient algorithm. The fuzzy Markovian segmentation is then performed automatically. The accuracy of the proposed method is quantitatively assessed on a digital phantom using an absolute average error and qualitatively tested on real MRI brain data. A comparison with the widely used fuzzy C-means algorithm is carried out to show numerous advantages of our method.     

Displaying of details in subvoxel accuracy

Under the volume segmentation in voxel space,a lot of details,some fine and thin objects,are ignored.In order to accurately display these details,this paper has developed a methodology for volume segmentation in subvoxel space.In the subvoxel space,most of the “bridges”between adjacent layers are broken down.Based on the subvoxel space,an automatic segmentation algorithm reserving details is discussed.After segmentation,volume data in subvoxel space are reduced to original voxel space.Thus,the details with width of only one or several voxels are extracted and displayed.     

组卷积轻量级脑肿瘤分割网络

目的 脑肿瘤是一种严重威胁人类健康的疾病。利用计算机辅助诊断进行脑肿瘤分割对于患者的预后和治疗具有重要的临床意义。3D卷积神经网络因具有空间特征提取充分、分割效果好等优点,广泛应用于脑肿瘤分割领域。但由于其存在显存占用量巨大、对硬件资源要求较高等问题,通常需要在网络结构中做出折衷,以牺牲精度或训练速度的方式来适应给定的内存预算。基于以上问题,提出一种轻量级分割算法。方法 使用组卷积来代替常规卷积以显著降低显存占用,并通过多纤单元与通道混合单元增强各组间信息交流。为充分利用多显卡协同计算的优势,使用跨卡同步批量归一化以缓解3D卷积神经网络因批量值过小所导致的训练效果差等问题。最后提出一种加权混合损失函数,提高分割准确性的同时加快模型收敛速度。结果 使用脑肿瘤公开数据集BraTS2018进行测试,本文算法在肿瘤整体区、肿瘤核心区和肿瘤增强区的平均Dice值分别可达90.67%、85.06%和80.41%,参数量和计算量分别为3.2 M和20.51 G,与当前脑肿瘤分割最优算法相比,其精度分别仅相差0.01%、0.96%和1.32%,但在参数量和计算量方面分别降低至对比算法的1/12和1/73。结论 本文算法通过加权混合损失函数来提高稀疏类分类错误对模型的惩罚,有效平衡不同分割难度类别的训练强度,本文算法可在保持较高精度的同时显著降低计算消耗,为临床医师进行脑肿瘤分割提供有力参考。     

串行处理卷积神经网络的海马子区分割

目的 海马体积很小,对比度极低,传统标记融合方法选用手工设计的特征模型,难以提取出适应性好、判别性强的特征。近年来,深度学习方法取得了极大成功,基于深度网络的方法已应用于医学图像分割中,但海马结构复杂,子区较多且体积差别较大,特别是CA2和CA3子区体积极小,常见的深度网络无法准确分割海马子区。为了解决这些问题,提出一种结合多尺度输入和串行处理神经网络的海马子区分割方法。方法 针对海马中体积差距较大的子区,设计两种不同的网络,结合多种尺度图像块信息,为小子区建立类别数量均衡的训练集,避免网络被极端化训练,最后,采用串行标记的方式对海马子区进行分割。结果 在Tail,SUB和PHG子区上的准确率达到了0.865,0.81,0.773,较现有的多图谱子区分割方法有较大提高,并且将体积较小子区CA2,CA3上的准确率分别提高了6%和9%。结论 该算法将基于卷积神经网络的分类方法引入到标记融合阶段,根据海马子区特殊的灰度及结构特点,设计两种针对性网络,实验证明,该算法能提取出适应性好、判别性强的特征,提高了分割准确率。     

An efficient classification and rendering method using tagged distance maps

Several high-speed volume rendering methods generate spatial data structures for speedup. Although they are useful for improving rendering speed, much time may be required to regenerate them. We propose an efficient classification and rendering method that supports fast classification. While original space-leaping needs to perform distance transformations for all voxels, our method modifies the values of some parts of the entire map by assigning predefined tag values when a voxels transparency is changed. The rendering algorithm is an extension of the space-leaping method and it determines the next sampling position by interpreting the values of those tagged voxels. This allows us to reclassify the volume quickly and to render datasets without loss of image quality.     

建立在图像识别基础上的体绘制加速方法

提出了一种新算法——IRVR（Image Recognition Volume Rendering）,该算法能大幅降低冗余数据,从而提升体绘制速度。IRVR算法首先利用交叉熵阈值分割法从三维数据集中将物体像素和背景像素识别出来,然后将迭代光线追踪方法和物体检测采样策略结合起来对原始三维数据集进行采样。接着运用快速迭代法对分类数据集进行采样,从而定位视线与原始数据集的交点。IRVR算法还应用了准确正规采样方法（例如,三线性插值、样条插值等）在体绘制过程中对原始数据集进行插值。经过实验得出的结论证明IRVR算法既能提高体绘制的速度,又可以保证体绘制图像的质量。     

巨核细胞图像分割方法的研究

随着现代医学研究的飞速发展,显微图像定量分析已经广泛应用到临床诊断、病理分析、癌变分级分类等越来越多的医学领域之内。巨核细胞是骨髓切片图像中个体较大的细胞,形状不规则,对其进行有效提取和分割对骨髓成份统计分析、血液疾病诊断等都有重要意义。该文根据数学形态学的知识,利用直方图势函数提取标记点,并将这些标记点作为种子点来对梯度图进行Watershed变换,进而实现骨髓切片中巨核细胞的有效分割。该方法是一种谱信息与空间信息相结合的分割方法,对分割结果与传统算法的对比分析表明,改进后的算法在分割的完整性和一致性上具有更好的效果。     

Research issues in volume visualization

Volume visualization is a method of extracting meaningful information from volumetric data sets through the use of interactive graphics and imaging. It addresses the representation, manipulation, and rendering of volumetric data sets, providing mechanisms for peering into structures and understanding their complexity and dynamics. Typically, the data set is represented as a 3D regular grid of volume elements (voxels) and stored in a volume buffer (also called a cubic frame buffer), which is a large 3D array of voxels. However, data is often defined at scattered or irregular locations that require using alternative representations and rendering algorithms. There are eight major research issues in volume visualization: volume graphics, volume rendering, transform coding of volume data, scattered data, enriching volumes with knowledge, segmentation, real-time rendering and parallelism, and special purpose hardware     

Probabilistic multiscale image segmentation

A method is presented to segment multidimensional images using a multiscale (hyperstack) approach with probabilistic linking. A hyperstack is a voxel-based multiscale data structure whose levels are constructed by convolving the original image with a Gaussian kernel of increasing width. Between voxels at adjacent scale levels, child-parent linkages are established according to a model-directed linkage scheme. In the resulting tree-like data structure, roots are formed to indicate the most plausible locations in scale space where segments in the original image are represented by a single voxel. The final segmentation is obtained by tracing back the linkages for all roots. The present paper deals with probabilistic (or multiparent) linking. The multiparent linkage structure is translated into a list of probabilities that are indicative of which voxels are partial volume voxels and to which extent. Probability maps are generated to visualize the progress of weak linkages in scale space when going from fine to coarser scale. It is demonstrated that probabilistic linking gives a significantly improved segmentation as compared with conventional (single-parent) linking     

一种肺部肿瘤CT图象序列的自动分割方法

肺部肿瘤序列图象的自动分割是计算机肺部肿瘤三维辅助诊断系统的关键技术之一,肿瘤与周围组织关系的复杂性造成分割困难.为了给医生提供准确的肺部肿瘤影像,运用纹理分析和径向基神经网络实现了肺部肿瘤CT图象序列的自动分割,并根据相邻层肿瘤图象灰度、位置的相关性,提出了一种自动获取多层肿瘤区域神经网络训练样本的阈值分割算法.该算法首先计算图象纹理统计参数,以组成特征矢量空间,然后利用自适应径向基神经网络对特征矢量进行分类来实现肿瘤序列图象的自动分割.实验结果表明,与基于灰度的区域增长法和基于梯度算子和形状算子的最优阈值的分割方法相比较,该方法不仅能充分利用肺部肿瘤序列图象的三维信息,还可最大限度地减少人工干预,且分割结果较好地表现了肿瘤形态特征,经临床医生评估,具有较好的临床指导价值.     

改进DRLSE的分步式肝脏及肿瘤分割方法

为了提升CT图像肝脏及肝脏肿瘤的分割精度,提出一种改进DRLSE的分步式肝脏及肿瘤分割方法。第一阶段：采用分步式分割方法对肝脏进行分割,（1）采用阈值处理、形态学方法、自适应区域生长方法进行肝脏的粗分割;（2）采用数学形态对分割结果进行优化,进行肝脏的细分割。第二阶段：构造参数梯度形态学和各向异性扩散滤波的距离正则化水平集演化（改进的DRLSE）模型进行肝脏肿瘤分割。实验利用3Dircadb数据集验证方法的有效性,计算了DICE、VOE、ASD和MSD指标评估分割的性能。实验结果表明该方法无需进行训练过程和统计模型的建立,对于复杂的形状和强度变化的CT图像分割效果尤为明显。由定量分析的数值结果显示,分割性能均优于比较算法,提高了分割准确率,具有较强的鲁棒性,为医生诊断和治疗肝癌提供帮助。     

On marching cubes

A characterization and classification of the isosurfaces of trilinear functions is presented. Based upon these results, a new algorithm for computing a triangular mesh approximation to isosurfaces for data given on a 3D rectilinear grid is presented. The original marching cubes algorithm is based upon linear interpolation along edges of the voxels. The asymptotic decider method is based upon bilinear interpolation on faces of the voxels. The algorithm of this paper carries this theme forward to using trilinear interpolation on the interior of voxels. The algorithm described here will produce a triangular mesh surface approximation to an isosurface which preserves the same connectivity/separation of vertices as given by the isosurface of trilinear interpolation.     

The unified extreme learning machines and discriminative random fields for automatic knee cartilage and meniscus segmentation from multi-contrast MR images

Segmenting articular cartilage and meniscus from magnetic resonance (MR) images is an essential task for the assessment of knee pathology. Most of the previous classification-based works for cartilage and meniscus segmentation only rely on independent labellings by a classifier, but do not consider the spatial context interaction. The labels of most image voxels are actually dependent upon their neighbours. In this study, we present an automatic knee segmentation system working on multi-contrast MR images where a novel classification model unifying an extreme learning machine (ELM)-based association potential and a discriminative random field (DRF)-based interaction potential is proposed. The DRF model introduces spatial dependencies between neighbouring voxels to the independent ELM classification. We exploit a rich set of features From multi-contrast MR images to train the proposed classification model and perform the loopy belief propagation for the inference. The proposed model is evaluated on multi-contrast MR datasets acquired from 11 subjects with results outperforming the independent classifiers in terms of segmentation accuracy of both cartilages and menisci.     

空间跳跃加速的GPU光线投射算法

光线投射算法是一种应用广泛的体绘制基本算法,能产生高质量的图像,但是时间复杂度较高。实现了一种基于图形处理器的单步光线投射算法,并在此基础上提出了一种基于空间跳跃技术的光线投射算法,以实现加速。采用八叉树组织体数据,利用空间跳跃有效地剔除体数据中对重建图像无贡献的部分,降低了硬件的负载。一个片段程序即可完成光线方向的生成、光线投射、空体素跳跃和光线终止等。实验结果表明,该算法对于内部包含大量空体素的体数据重建能起到明显的加速作用。