Combination Strategies in Multi-Atlas Image Segmentation: Application to Brain MR Data

It has been shown that employing multiple atlas images improves segmentation accuracy in atlas-based medical image segmentation. Each atlas image is registered to the target image independently and the calculated transformation is applied to the segmentation of the atlas image to obtain a segmented version of the target image. Several independent candidate segmentations result from the process, which must be somehow combined into a single final segmentation. Majority voting is the generally used rule to fuse the segmentations, but more sophisticated methods have also been proposed. In this paper, we show that the use of global weights to ponderate candidate segmentations has a major limitation. As a means to improve segmentation accuracy, we propose the generalized local weighting voting method. Namely, the fusion weights adapt voxel-by-voxel according to a local estimation of segmentation performance. Using digital phantoms and MR images of the human brain, we demonstrate that the performance of each combination technique depends on the gray level contrast characteristics of the segmented region, and that no fusion method yields better results than the others for all the regions. In particular, we show that local combination strategies outperform global methods in segmenting high-contrast structures, while global techniques are less sensitive to noise when contrast between neighboring structures is low. We conclude that, in order to achieve the highest overall segmentation accuracy, the best combination method for each particular structure must be selected.      

Performance-based classifier combination in atlas-based image segmentation using expectation-maximization parameter estimation

It is well known in the pattern recognition community that the accuracy of classifications obtained by combining decisions made by independent classifiers can be substantially higher than the accuracy of the individual classifiers. We have previously shown this to be true for atlas-based segmentation of biomedical images. The conventional method for combining individual classifiers weights each classifier equally (vote or sum rule fusion). In this paper, we propose two methods that estimate the performances of the individual classifiers and combine the individual classifiers by weighting them according to their estimated performance. The two methods are multiclass extensions of an expectation-maximization (EM) algorithm for ground truth estimation of binary classification based on decisions of multiple experts (Warfield et al., 2004). The first method performs parameter estimation independently for each class with a subsequent integration step. The second method considers all classes simultaneously. We demonstrate the efficacy of these performance-based fusion methods by applying them to atlas-based segmentations of three-dimensional confocal microscopy images of bee brains. In atlas-based image segmentation, multiple classifiers arise naturally by applying different registration methods to the same atlas, or the same registration method to different atlases, or both. We perform a validation study designed to quantify the success of classifier combination methods in atlas-based segmentation. By applying random deformations, a given ground truth atlas is transformed into multiple segmentations that could result from imperfect registrations of an image to multiple atlas images. In a second evaluation study, multiple actual atlas-based segmentations are combined and their accuracies computed by comparing them to a manual segmentation. We demonstrate in both evaluation studies that segmentations produced by combining multiple individual registration-based segmentations are more accurate for the two classifier fusion methods we propose, which weight the individual classifiers according to their EM-based performance estimates, than for simple sum rule fusion, which weights each classifier equally.     

Improved labeling of subcortical brain structures in atlas-based segmentation of magnetic resonance images

Precise labeling of subcortical structures plays a key role in functional neurosurgical applications. Labels from an atlas image are propagated to a patient image using atlas-based segmentation. Atlas-based segmentation is highly dependent on the registration framework used to guide the atlas label propagation. This paper focuses on atlas-based segmentation of subcortical brain structures and the effect of different registration methods on the generated subcortical labels. A single-step and three two-step registration methods appearing in the literature based on affine and deformable registration algorithms in the ANTS and FSL algorithms are considered. Experiments are carried out with two atlas databases of IBSR and LPBA40. Six segmentation metrics consisting of Dice overlap, relative volume error, false positive, false negative, surface distance, and spatial extent are used for evaluation. Segmentation results are reported individually and as averages for nine subcortical brain structures. Based on two statistical tests, the results are ranked. In general, among four different registration strategies investigated in this paper, a two-step registration consisting of an initial affine registration followed by a deformable registration applied to subcortical structures provides superior segmentation outcomes. This method can be used to provide an improved labeling of the subcortical brain structures in MRIs for different applications.     

基于GAN的ISAR图像语义分割方法

逆合成孔径雷达（ISAR）成像技术能够对空间目标进行远距离成像,刻画目标的外形、结构和尺寸等信息。ISAR图像语义分割能够获取目标的感兴趣区域,是ISAR图像解译的重要技术支撑,具有非常重要的研究价值。由于ISAR图像表征性较差,图像中散射点的不连续和强散射点存在的旁瓣效应使得人工精准标注十分困难,基于交叉熵损失的传统深度学习语义分割方法在语义标注不精准情况下无法保证分割性能的稳健。针对这一问题,提出了一种基于生成对抗网络（GAN）的ISAR图像语义分割方法,采用对抗学习思想学习ISAR图像分布到其语义分割图像分布的映射关系,同时通过构建分割图像的局部信息和全局信息来保证语义分割的精度。基于仿真卫星目标ISAR图像数据集的实验结果证明,本文方法能够取得较好的语义分割结果,且在语义标注不够精准的情况下模型更稳健。     

基于多参数配准模型的脑核磁影像分割算法

配准技术在基于多图谱的分割方法中能有效地将医学图谱的先验知识融入分割过程,再结合以高效的标记融合算法,最终实现精确地自动分割.针对图谱配准的较大误差及其对标记融合的重要影响,本文建立了一种新的概率图模型框架并以此提出了基于多参数配准模型的分割算法,将此方法与高效的标记融合算法相结合,可以提高目标图像中特定组织区域的分割精度,更使其在少量图谱分割的情形下具有重要应用.首先,使用多种配准参数对所有目标图像进行配准;然后,分别采用不同的算法对配准图像进行灰度融合和标记融合,实现训练图像的重构过程;最后,利用高效的标记融合算法对重构后的图像进行融合得到最终精确的分割结果.实验结果表明该方法均优于本文其他分割算法,能够有效提升脑部组织分割精度.     

Multiple abdominal organ segmentation: an atlas-based fuzzy connectedness approach.

Organ segmentation is an important step in various medical image applications. In this paper, a presegmented atlas is incorporated into the fuzzy connectedness (FC) framework for automatic segmentation of abdominal organs. First, the atlas is registered onto the subject to provide an initial segmentation. Then, a novel method is applied to estimate the necessary FC parameters such as organ intensity features, seeds, and optimal FC threshold automatically and subject adaptively. In order to overcome the intensity overlapping between the neighboring organs, a shape modification approach based on Euclidean distance and watershed segmentation is used. This atlas-based segmentation method has been tested on some abdominal CT and MRI images from Chinese patients. Experimental results indicate the validity of this segmentation method for various image modalities.     

Segmentation of overlapping cervical nuclei based on the identification

Image segmentation directly determines the performance of automatic screening technique. However, there are overlapping nuclei in nuclei images. It raises a challenge to nuclei segmentation. To solve the problem, a segmentation method of overlapping cervical nuclei based on the identification is proposed. This method consists of three stages: classifier training, recognition and fine segmentation. In the classifier training, feature selection and classifier selection are used to obtain a classifier with high recognition rate. In the recognition, the outputs of the rough segmentation are classified and processed according to their labels. In the fine segmentation, the severely overlapping nuclei are further segmented based on the prior knowledge provided by the recognition. Experiments show that this method can accurately segment overlapping nuclei.     

基于阶梯结构的U-Net结肠息肉分割算法

结肠息肉的精确分割对结直肠癌的诊断和治疗具有重要意义,目前的分割方法普遍存在有伪影、分割精度低等问题.该文提出一种基于阶梯结构的U-Net结肠息肉分割算法(SU-Net),使用U-Net的U型结构,利用Kronecker乘积来扩展标准空洞卷积核,构成Kronecker空洞卷积下采样有效扩大感受野,弥补传统空洞卷积容易丢...     

Automated Pulmonary Fibrosis Segmentation Using a 3D Multi-Scale Convolutional Encoder-Decoder Approach in Thoracic CT for the Rhesus Macaque with Radiation-Induced Lung Damage

To develop an automated pulmonary fibrosis (PF) segmentation methodology using a 3D multi-scale convolutional encoder-decoder approach following the robust atlas-based active volume model in thoracic CT for Rhesus Macaques with radiation-induced lung damage. 152 thoracic computed tomography scans of Rhesus Macaques with radiation-induced lung damage were collected. The 3D input data are randomly augmented with the Gaussian blurring when applying the 3D multi-scale convolutional encoder-decoder (3D MSCED) segmentation method.PF in each scan was manually segmented in which 70% scans were used as training data, 20% scans were used as validation data, and 10% scans were used as testing data. The performance of the method is assessed based on a10-fold cross validation method. The workflow of the proposed method has two parts. First, the compromised lung volume with acute radiation-induced PF was segmented using a robust atlas-based active volume model. Next, a 3D multi-scale convolutional encoder-decoder segmentation method was developed which merged the higher spatial information from low-level features with the high-level object knowledge encoded in upper network layers. It included a bottom-up feed-forward convolutional neural network and a top-down learning mask refinement process. The quantitative results of our segmentation method achieved mean Dice score of (0.769, 0.853), mean accuracy of (0.996, 0.999), and mean relative error of (0.302, 0.512) with 95% confidence interval. The qualitative and quantitative comparisons show that our proposed method can achieve better segmentation accuracy with less variance in testing data. This method was extensively validated in NHP datasets. The results demonstrated that the approach is more robust relative to PF than other methods. It is a general framework which can easily be applied to segmentation other lung lesions.

     

Segmentation by fusion of histogram-based k-means clusters in different color spaces.

This paper presents a new, simple, and efficient segmentation approach, based on a fusion procedure which aims at combining several segmentation maps associated to simpler partition models in order to finally get a more reliable and accurate segmentation result. The different label fields to be fused in our application are given by the same and simple (K-means based) clustering technique on an input image expressed in different color spaces. Our fusion strategy aims at combining these segmentation maps with a final clustering procedure using as input features, the local histogram of the class labels, previously estimated and associated to each site and for all these initial partitions. This fusion framework remains simple to implement, fast, general enough to be applied to various computer vision applications (e.g., motion detection and segmentation), and has been successfully applied on the Berkeley image database. The experiments herein reported in this paper illustrate the potential of this approach compared to the state-of-the-art segmentation methods recently proposed in the literature.     

基于YCbCr空间对大壁虎脑图谱的分割

确定大壁虎的各种运动行为与脑内控制其运动的具体的皮层组织对应关系是壁虎机器人的制作关键。通过对其脑图谱的制备,可以观察到脑内皮层组织的丰富信息。针对HE染色的大壁虎脑图谱自身的独特性,经过推敲,提出一种基于YCbCr空间的对图谱进行分割的方法。该方法将图谱由RGB空间转换为YCbCr空间,将亮度分量独立出来,不仅解决了拍摄过程中由于光线改变所带来的问题,而且阈值分明,分割彻底,鲁棒性好。其结果表明基于YCbCr空间的分割方法是分割效果非常好的方法。     

基于多特征图金字塔融合深度网络的遥感图像语义分割

在遥感图像语义分割中,利用多元数据(如高程信息)进行辅助是一个研究重点。现有的基于多元数据的分割方法通常直接将多元数据作为模型的多特征输入,未能充分利用多元数据的多层次特征,此外,遥感图像中目标尺寸大小不一,对于一些中小型目标,如车辆、房屋等,难以做到精细化分割。针对以上问题,提出一种多特征图金字塔融合深度网络(MFPNet),该模型利用光学遥感图像和高程数据作为输入,提取图像的多层次特征,然后针对不同层次的特征,分别引入金字塔池化结构,提取图像的多尺度特征,最后,设计了一种多层次、多尺度特征融合策略,综合利用多元数据的特征信息,实现遥感图像的精细化分割。基于Vaihingen数据集设计了相应的对比实验,实验结果证明了所提方法的有效性。     

Automatic 3-D segmentation of internal structures of the head in MR images using a combination of similarity and free-form transformations: Part II, validation on severely atrophied brains

Studies aimed at quantifying neuroanatomical differences between populations require the volume measurements of individual brain structures. If the study contains a large number of images, manual segmentation is not practical. This study tests the hypothesis that a fully automatic, atlas-based segmentation method can be used to quantify atrophy indexes derived from the brain and cerebellum volumes in normal subjects and chronic alcoholics. This is accomplished by registering an atlas volume with a subject volume, first using a global transformation, and then improving the registration using a local transformation. Segmented structures in the atlas volume are then mapped to the corresponding structures in the subject volume using the combined global and local transformations. This technique has been applied to seven normal and seven alcoholic subjects. Three magnetic resonance volumes were obtained for each subject and each volume was segmented automatically, using the atlas-based method. Accuracy was assessed by manually segmenting regions and measuring the similarity between corresponding regions obtained automatically. Repeatability was determined by comparing volume measurements of segmented structures from each acquisition of the same subject. Results demonstrate that the method is accurate, that the results are repeatable, and that it can provide a method for automatic quantification of brain atrophy, even when the degree of atrophy is large.     

基于红外图像帧关联的自动阈值分割方法

针对目前单帧图像阈值分割中分割易受突变影响、目 标背景分割不明显以及分割效果较差等问题,提出了一种基于红外图像帧关 联的自动阈值分割方法。该方法利用自动阈值分割法简单分割单帧图像,然 后根据图像帧关联信息对图像进行分组处理,再对每帧图像进行权重分配,最 终确定每帧图像的分割阈值,以提高分割的抗干扰性,改善分割效果。通过理 论分析和实验仿真验证了该算法的有效性和可行性,并将其与其他算法进行了 对比实验。实验结果表明,本文提出的分割算法的抗干扰性较强,能够将目标 图像从背景中清晰地分割出来,具有更好的分割效果和更强的应用性。     

Segmentation by Fusion of Histogram-Based $K$-Means Clusters in Different Color Spaces

This paper presents a new, simple, and efficient segmentation approach, based on a fusion procedure which aims at combining several segmentation maps associated to simpler partition models in order to finally get a more reliable and accurate segmentation result. The different label fields to be fused in our application are given by the same and simple (K-means based) clustering technique on an input image expressed in different color spaces. Our fusion strategy aims at combining these segmentation maps with a final clustering procedure using as input features, the local histogram of the class labels, previously estimated and associated to each site and for all these initial partitions. This fusion framework remains simple to implement, fast, general enough to be applied to various computer vision applications (e.g., motion detection and segmentation), and has been successfully applied on the Berkeley image database. The experiments herein reported in this paper illustrate the potential of this approach compared to the state-of-the-art segmentation methods recently proposed in the literature.     

并行MCMC算法的SAR影像分割

MCMC(Markov Chain Monte Carlo, MCMC)方法采用顺序改变表征像素类属性的标号变量值会导致算法运算时间长、收敛速度慢等问题。为此,本文提出并行化改变像素标号值的MCMC方案,在贝叶斯推理框架下,依据高斯分布及MRF(Markov Random Field, MRF)模型建立SAR(Synthetic Aperture Radar, SAR)影像分割模型,设计实现基于多线程的并行采样方案;为了解决MRF标号场中邻域像素标号相关性问题,提出独立的像素并行采样的准则;同时,限制并行线程的数量,以保证采样的随机性。运用传统的串行算法和提出的并行算法对模拟和真实SAR影像进行影像分割实验;定性和定量的时间和精度评价结果表明:该方案在不影响分割精度的前提下大幅缩短影像分割时间,提高了效率。      

基于表征转换机的SAR图像目标分割方法

针对SAR(Synthetic Aperture Radar)图像中的目标分割问题,由于目标与杂波空间模式(像素强度和分布)不同,通过分析图像空间模式的方式可达到分辨目标和杂波并分割目标的目的。该文基于表征转换机理论提出一种有效的SAR图像目标分割方法,该算法分析SAR图像中的空间模式,计算其与参考杂波图像的相似程度,最后将与参考杂波相似程度较高的部分消除以达到分割目标的目的,并在衡量相似度部分使用基于累积直方图的自动阈值选取办法。仿真和实测数据的实验验证了此算法的有效性。      

Automatic tractography segmentation using a high-dimensional white matter atlas

We propose a new white matter atlas creation method that learns a model of the common white matter structures present in a group of subjects. We demonstrate that our atlas creation method, which is based on group spectral clustering of tractography, discovers structures corresponding to expected white matter anatomy such as the corpus callosum, uncinate fasciculus, cingulum bundles, arcuate fasciculus, and corona radiata. The white matter clusters are augmented with expert anatomical labels and stored in a new type of atlas that we call a high-dimensional white matter atlas. We then show how to perform automatic segmentation of tractography from novel subjects by extending the spectral clustering solution, stored in the atlas, using the Nystrom method. We present results regarding the stability of our method and parameter choices. Finally we give results from an atlas creation and automatic segmentation experiment. We demonstrate that our automatic tractography segmentation identifies corresponding white matter regions across hemispheres and across subjects, enabling group comparison of white matter anatomy.     

Multiscale modeling for image analysis of brain tumor studies

Image-based modeling of tumor growth combines methods from cancer simulation and medical imaging. In this context, we present a novel approach to adapt a healthy brain atlas to MR images of tumor patients. In order to establish correspondence between a healthy atlas and a pathologic patient image, tumor growth modeling in combination with registration algorithms is employed. In a first step, the tumor is grown in the atlas based on a new multiscale, multiphysics model including growth simulation from the cellular level up to the biomechanical level, accounting for cell proliferation and tissue deformations. Large-scale deformations are handled with an Eulerian approach for finite element computations, which can operate directly on the image voxel mesh. Subsequently, dense correspondence between the modified atlas and patient image is established using nonrigid registration. The method offers opportunities in atlas-based segmentation of tumor-bearing brain images as well as for improved patient-specific simulation and prognosis of tumor progression.     

Atlas renormalization for improved brain MR image segmentation across scanner platforms

Atlas-based approaches have demonstrated the ability to automatically identify detailed brain structures from 3-D magnetic resonance (MR) brain images. Unfortunately, the accuracy of this type of method often degrades when processing data acquired on a different scanner platform or pulse sequence than the data used for the atlas training. In this paper, we improve the performance of an atlas-based whole brain segmentation method by introducing an intensity renormalization procedure that automatically adjusts the prior atlas intensity model to new input data. Validation using manually labeled test datasets has shown that the new procedure improves the segmentation accuracy (as measured by the Dice coefficient) by 10% or more for several structures including hippocampus, amygdala, caudate, and pallidum. The results verify that this new procedure reduces the sensitivity of the whole brain segmentation method to changes in scanner platforms and improves its accuracy and robustness, which can thus facilitate multicenter or multisite neuroanatomical imaging studies.