A robust method for extraction and automatic segmentation of brain images

A new protocol is introduced for brain extraction and automatic tissue segmentation of MR images. For the brain extraction algorithm, proton density and T2-weighted images are used to generate a brain mask encompassing the full intracranial cavity. Segmentation of brain tissues into gray matter (GM), white matter (WM), and cerebral spinal fluid (CSF) is accomplished on a T1-weighted image after applying the brain mask. The fully automatic segmentation algorithm is histogram-based and uses the Expectation Maximization algorithm to model a four-Gaussian mixture for both global and local histograms. The means of the local Gaussians for GM, WM, and CSF are used to set local thresholds for tissue classification. Reproducibility of the extraction procedure was excellent, with average variation in intracranial capacity (TIC) of 0.13 and 0.66% TIC in 12 healthy normal and 33 Alzheimer brains, respectively. Repeatability of the segmentation algorithm, tested on healthy normal images, indicated scan-rescan differences in global tissue volumes of less than 0.30% TIC. Reproducibility at the regional level was established by comparing segmentation results within the 12 major Talairach subdivisions. Accuracy of the algorithm was tested on a digital brain phantom, and errors were less than 1% of the phantom volume. Maximal Type I and Type II classification errors were low, ranging between 2.2 and 4.3% of phantom volume. The algorithm was also insensitive to variation in parameter initialization values. The protocol is robust, fast, and its success in segmenting normal as well as diseased brains makes it an attractive clinical application.     

三维多模态医学图像自动配准方法研究

提出一种综合应用图像分割与互信息的医学图像自动配准方法.首先采用门限法和数学形态学方法进行预处理,再用k-means方法进行分割,之后采用基于互信息的Powell优化方法配准.将该方法用于磁共振图像(MRI)和正电子发射断层扫描(PET)临床医学图像配准,得到较满意的效果.     

Dynamic-weighting hierarchical segmentation network for medical images

Automatic medical image segmentation plays a crucial role in many medical image analysis applications, such as disease diagnosis and prognosis. Despite the extensive progress of existing deep learning based models for medical image segmentation, they focus on extracting accurate features by designing novel network structures and solely utilize fully connected (FC) layer for pixel-level classification. Considering the insufficient capability of FC layer to encode the extracted diverse feature representations, we propose a Hierarchical Segmentation (HieraSeg) Network for medical image segmentation and devise a Hierarchical Fully Connected (HFC) layer. Specifically, it consists of three classifiers and decouples each category into several subcategories by introducing multiple weight vectors to denote the diverse characteristics in each category. A subcategory-level and a category-level learning schemes are then designed to explicitly enforce the discrepant subcategories and automatically capture the most representative characteristics. Hence, the HFC layer can fit the variant characteristics so as to derive an accurate decision boundary. To enhance the robustness of HieraSeg Network with the variability of lesions, we further propose a Dynamic-Weighting HieraSeg (DW-HieraSeg) Network, which introduces an Image-level Weight Net (IWN) and a Pixel-level Weight Net (PWN) to learn data-driven curriculum. Through progressively incorporating informative images and pixels in an easy-to-hard manner, DW-HieraSeg Network is able to eliminate local optimums and accelerate the training process. Additionally, a class balanced loss is proposed to constrain the PWN for preventing the overfitting problem in minority category. Comprehensive experiments on three benchmark datasets, EndoScene, ISIC and Decathlon, show our newly proposed HieraSeg and DW-HieraSeg Networks achieve state-of-the-art performance, which clearly demonstrates the effectiveness of the proposed approaches for medical image segmentation.     

An SPCNN-GVF-based approach for the automatic segmentation of left ventricle in cardiac cine MR images

Purpose

Accurate segmentation of left ventricle (LV) is essential for the cardiac function analysis. However, it is labor intensive and time consuming for radiologists to delineate LV boundary manually. In this paper, we present a novel self-correcting framework for the fully automatic LV segmentation.

Methods

Firstly, a time-domain method is designed to extract a rectangular region of interest around the heart. Then, the simplified pulse-coupled neural network (SPCNN) is employed to locate the LV cavity. Different from the existing approaches, SPCNN can realize the self-correcting segmentation due to its parameter controllability. Subsequently, the post-processing based on the maximum gradient searching is proposed to obtain the accurate endocardium. Finally, a new external force based on the shape similarity is defined and integrated into the gradient vector flow (GVF) snake with the balloon force to segment the epicardium.

Results

We obtain encouraging segmentation results tested on the database provided by MICCAI 2009. The average percentage of good contours is 92.26 %, the average perpendicular distance is 2.38 mm, and the overlapping dice metric is 0.89. Besides, the experiment results show good correlations between the automatic segmentation and the manual delineation (for the LV ejection fraction and the LV myocardial mass, the correlation coefficients R are 0.9683 and 0.9278, respectively).

Conclusion

We propose an effective and fast method combing the SPCNN and the improved GVF for the automatic segmentation of LV.

     

Interaction in the segmentation of medical images: a survey

Segmentation of the object of interest is a difficult step in the analysis of digital images. Fully automatic methods sometimes fail, producing incorrect results and requiring the intervention of a human operator. This is often true in medical applications, where image segmentation is particularly difficult due to restrictions imposed by image acquisition, pathology and biological variation. In this paper we present an early review of the largely unknown territory of human-computer interaction in image segmentation. The purpose is to identify patterns in the use of interaction and to develop qualitative criteria to evaluate interactive segmentation methods. We discuss existing interactive methods with respect to the following aspects: the type of information provided by the user, how this information affects the computational part, and the purpose of interaction in the segmentation process. The discussion is based on the potential impact of each strategy on the accuracy, repeatability and interaction efficiency. Among others, these are important aspects to characterise and understand the implications of interaction to the results generated by an interactive segmentation method. This survey is focused on medical imaging, however similar patterns are expected to hold for other applications as well.     

一种鲁棒的骨龄X射线平片自动轮廓提取方法

背景:骨龄X射线平片具有不均匀和复杂性,因而在骨龄自动评价的研究中,手掌轮廓提取的结果往往不理想.目的:采用计算机自动提取手掌轮廓,为骨龄自动评价中图像预处理阶段的研究奠定重要基础.为了解除骨龄评定带来的主观性和不确定性,提出用计算机进行自动评价.方法:在仔细分析骨龄X射线平片的基础上,提出了一种对图像背景的可行有效子采样点方案,并提出用二元三次线性回归方法来模拟图像背景,通过形态学以及二进制标记等～系列操作,最后成功提取出手掌轮廓.结果与结论:采用异常点移除和回归方法相结合来提取轮廓,采取固定阀值,不受阀值选取的困扰,并具有鲁棒性.大量的实验结果表明了该方法提取的手掌轮廓成功率在93%以上,能完全应用于骨龄自动识别的后续研究工作.     

An end-to-end approach to segmentation in medical images with CNN and posterior-CRF

Conditional Random Fields (CRFs) are often used to improve the output of an initial segmentation model, such as a convolutional neural network (CNN). Conventional CRF approaches in medical imaging use manually defined features, such as intensity to improve appearance similarity or location to improve spatial coherence. These features work well for some tasks, but can fail for others. For example, in medical image segmentation applications where different anatomical structures can have similar intensity values, an intensity-based CRF may produce incorrect results. As an alternative, we propose Posterior-CRF, an end-to-end segmentation method that uses CNN-learned features in a CRF and optimizes the CRF and CNN parameters concurrently. We validate our method on three medical image segmentation tasks: aorta and pulmonary artery segmentation in non-contrast CT, white matter hyperintensities segmentation in multi-modal MRI, and ischemic stroke lesion segmentation in multi-modal MRI. We compare this with the state-of-the-art CNN-CRF methods. In all applications, our proposed method outperforms the existing methods in terms of Dice coefficient, average volume difference, and lesion-wise F1 score.     

FCN-based approach for the automatic segmentation of bone surfaces in ultrasound images

Purpose

A new algorithm, based on fully convolutional networks (FCN), is proposed for the automatic localization of the bone interface in ultrasound (US) images. The aim of this paper is to compare and validate this method with (1) a manual segmentation and (2) a state-of-the-art method called confidence in phase symmetry (CPS).

Methods

The dataset used for this study was composed of 1738 US images collected from three volunteers and manually delineated by three experts. The inter- and intra-observer variabilities of this manual delineation were assessed. Images having annotations with an inter-observer variability higher than a confidence threshold were rejected, resulting in 1287 images. Both FCN-based and CPS approaches were studied and compared to the average inter-observer segmentation according to six criteria: recall, precision, F1 score, accuracy, specificity and root-mean-square error (RMSE).

Results

The intra- and inter-observer variabilities were inferior to 1 mm for 90% of manual annotations. The RMSE was 1.32?±?3.70  mm and 5.00?±?7.70 mm for, respectively, the FCN-based approach and the CPS algorithm. The mean recall, precision, F1 score, accuracy and specificity were, respectively, 62%, 64%, 57%, 80% and 83% for the FCN-based approach and 66%, 34%, 41%, 52% and 43% for the CPS algorithm.

Conclusion

The FCN-based approach outperforms the CPS algorithm, and the obtained RMSE is similar to the manual segmentation uncertainty.

     

Semantic instance segmentation with discriminative deep supervision for medical images

Semantic instance segmentation is crucial for many medical image analysis applications, including computational pathology and automated radiation therapy. Existing methods for this task can be roughly classified into two categories: (1) proposal-based methods and (2) proposal-free methods. However, in medical images, the irregular shape-variations and crowding instances (e.g., nuclei and cells) make it hard for the proposal-based methods to achieve robust instance localization. On the other hand, ambiguous boundaries caused by the low-contrast nature of medical images (e.g., CT images) challenge the accuracy of the proposal-free methods. To tackle these issues, we propose a proposal-free segmentation network with discriminative deep supervision (DDS), which at the same time allows us to gain the power of the proposal-based method. The DDS module is interleaved with a carefully designed proposal-free segmentation backbone in our network. Consequently, the features learned by the backbone network become more sensitive to instance localization. Also, with the proposed DDS module, robust pixel-wise instance-level cues (especially structural information) are introduced for semantic segmentation. Extensive experiments on three datasets, i.e., a nuclei dataset, a pelvic CT image dataset, and a synthetic dataset, demonstrate the superior performance of the proposed algorithm compared to the previous works.     

Morphology-driven automatic segmentation of MR images of the neonatal brain

The segmentation of MR images of the neonatal brain is an essential step in the study and evaluation of infant brain development. State-of-the-art methods for adult brain MRI segmentation are not applicable to the neonatal brain, due to large differences in structure and tissue properties between newborn and adult brains. Existing newborn brain MRI segmentation methods either rely on manual interaction or require the use of atlases or templates, which unavoidably introduces a bias of the results towards the population that was used to derive the atlases. We propose a different approach for the segmentation of neonatal brain MRI, based on the infusion of high-level brain morphology knowledge, regarding relative tissue location, connectivity and structure. Our method does not require manual interaction, or the use of an atlas, and the generality of its priors makes it applicable to different neonatal populations, while avoiding atlas-related bias. The proposed algorithm segments the brain both globally (intracranial cavity, cerebellum, brainstem and the two hemispheres) and at tissue level (cortical and subcortical gray matter, myelinated and unmyelinated white matter, and cerebrospinal fluid). We validate our algorithm through visual inspection by medical experts, as well as by quantitative comparisons that demonstrate good agreement with expert manual segmentations. The algorithm’s robustness is verified by testing on variable quality images acquired on different machines, and on subjects with variable anatomy (enlarged ventricles, preterm- vs. term-born).     

An algorithm for calculi segmentation on ureteroscopic images

Purpose  

The purpose of the study is to develop an algorithm for the segmentation of renal calculi on ureteroscopic images. In fact, renal calculi are common source of urological obstruction, and laser lithotripsy during ureteroscopy is a possible therapy. A laser-based system to sweep the calculus surface and vaporize it was developed to automate a very tedious manual task. The distal tip of the ureteroscope is directed using image guidance, and this operation is not possible without an efficient segmentation of renal calculi on the ureteroscopic images.     

Feasibility of the soft attention-based models for automatic segmentation of OCT kidney images

Clinically, optical coherence tomography (OCT) has been utilized to obtain the images of the kidney’s proximal convoluted tubules (PCTs), which can be used to quantify the morphometric parameters such as tubular density and diameter. Such parameters are useful for evaluating the status of the donor kidney for transplant. Quantifying PCTs from OCT images by human readers is a time-consuming and tedious process. Despite the fact that conventional deep learning models such as conventional neural networks (CNNs) have achieved great success in the automatic segmentation of kidney OCT images, gaps remain regarding the segmentation accuracy and reliability. Attention-based deep learning model has benefits over regular CNNs as it is intended to focus on the relevant part of the image and extract features for those regions. This paper aims at developing an Attention-based UNET model for automatic image analysis, pattern recognition, and segmentation of kidney OCT images. We evaluated five methods including the Residual-Attention-UNET, Attention-UNET, standard UNET, Residual UNET, and fully convolutional neural network using 14403 OCT images from 169 transplant kidneys for training and testing. Our results show that Residual-Attention-UNET outperformed the other four methods in segmentation by showing the highest values of all the six metrics including dice score (0.81 ± 0.01), intersection over union (IOU, 0.83 ± 0.02), specificity (0.84 ± 0.02), recall (0.82 ± 0.03), precision (0.81 ± 0.01), and accuracy (0.98 ± 0.08). Our results also show that the performance of the Residual-Attention-UNET is equivalent to the human manual segmentation (dice score = 0.84 ± 0.05). Residual-Attention-UNET and Attention-UNET also demonstrated good performance when trained on a small dataset (3456 images) whereas the performance of the other three methods dropped dramatically. In conclusion, our results suggested that the soft Attention-based models and specifically Residual-Attention-UNET are powerful and reliable methods for tubule lumen identification and segmentation and can help clinical evaluation of transplant kidney viability as fast and accurate as possible.     

A hybrid approach for segmentation of anatomic structures in medical images

Objective

We propose a hybrid interactive approach for the segmentation of anatomic structures in medical images with higher accuracy at lower user interaction cost.

Materials and methods

Eighteen brain MR scans from the Internet Brain Segmentation Repository are used for brain structure segmentation. A MR scan and a CT scan of an old female are used for orbital structure segmentation. The proposed approach combines shape-based interpolation, radial basis function (RBF)-based warping and model-based segmentation. With this approach, to segment a structure in a 3D image, we first delineate the structure in several slices using interactive methods, and then use shape-based interpolation to automatically generate an initial 3D model of the structure from the segmented slices. To refine the initial model, we specify a set of additional points on the structure boundary in the image, and use a RBF to warp the model so that it passes the specified points. Finally, we adopt a point-anchored active surface approach to further deform the model for a better fitting of the model with its corresponding structure in image.

Results

Two brain structures and 15 orbital structures are segmented. For each structure, it needs only to semi- automatically segment three to five 2D slices and specify two to nine additional points on the structure boundary. The time cost for each structure is about 1–3 min. The overlap ratio of the segmentation results and the ground truth is higher than 96%.

Conclusion

The proposed method for the segmentation of anatomic structure achieved higher accuracy at lower user interaction cost, and therefore promising in many applications such as surgery planning and simulation, atlas construction, and morphometric analysis of anatomic structures.     

MAP MRF joint segmentation and registration of medical images

The problems of segmentation and registration are traditionally approached individually, yet the accuracy of one is of great importance in influencing the success of the other. In this paper, we aim to show that more accurate and robust results may be obtained through seeking a joint solution to these linked processes. The outlined approach applies Markov random fields in the solution of a maximum a posteriori model of segmentation and registration. The approach is applied to synthetic and real MRI data.     

A review of automatic mass detection and segmentation in mammographic images

The aim of this paper is to review existing approaches to the automatic detection and segmentation of masses in mammographic images, highlighting the key-points and main differences between the used strategies. The key objective is to point out the advantages and disadvantages of the various approaches. In contrast with other reviews which only describe and compare different approaches qualitatively, this review also provides a quantitative comparison. The performance of seven mass detection methods is compared using two different mammographic databases: a public digitised database and a local full-field digital database. The results are given in terms of Receiver Operating Characteristic (ROC) and Free-response Receiver Operating Characteristic (FROC) analysis.     

A geometric method for the detection and correction of segmentation leaks of anatomical structures in volumetric medical images

Purpose

Patient-specific models of anatomical structures and pathologies generated from volumetric medical images play an increasingly central role in many aspects of patient care. A key task in generating these models is the segmentation of anatomical structures and pathologies of interest. Although numerous segmentation methods are available, they often produce erroneous delineations that require time-consuming modifications.

Methods

   We present a new geometry-based algorithm for the reliable detection and correction of segmentation errors in volumetric medical images. The method is applicable to anatomical structures consisting of a few 3D star-shaped components. First, it detects segmentation errors by casting rays from the initial segmentation interior to its outer surface. It then classifies the segmentation surface into correct and erroneous regions by minimizing an energy functional that incorporates first- and second-order properties of the rays lengths. Finally, it corrects the segmentation errors by computing new locations for the erroneous surface points by Laplace deformation so that the new surface has maximum smoothness with respect to the rays-length gradient magnitude.

Results

   Our evaluation on initial segmentations of 16 abdominal aortic aneurysm and 12 lung tumors in CT scans obtained by both adaptive region-growing and active contours level-set segmentation improved the volumetric overlap error by 66 and 70.5 % respectively, with respect to the ground-truth.

Conclusions

   The advantages of our method are that it is independent of the initial segmentation algorithm that covers a variety of anatomical structures and pathologies, that it does not require a shape prior, and that it requires minimal user interaction.

     

Object-based analysis of CT images for automatic detection and segmentation of hypodense liver lesions

Purpose  

Hypodense liver lesions are commonly detected in CT, so their segmentation and characterization are essential for diagnosis and treatment. Methods for automatic detection and segmentation of liver lesions were developed to support this task.     

一种数字人脑部切片图像分割新方法

目的 提出一种人脑切片图像自动分割算法,以克服现有的方法对大量人工参与的依赖.方法 针对人脑切片图像的特征,提出一种基于区域生长的灰度直方图阈值化分割算法.首先通过区域生长过程对图像进行初始的粗分割,再用直方图阈值化方法进行二次细分割提取目标区域.结果 采用此方法准确有效地分割出了大脑白质和大脑皮质.结论 此算法结合切片图像的全局信息和局部信息应用于分割,是一种比较好的分割方法.     

Unified generative adversarial networks for multimodal segmentation from unpaired 3D medical images

To fully define the target objects of interest in clinical diagnosis, many deep convolution neural networks (CNNs) use multimodal paired registered images as inputs for segmentation tasks. However, these paired images are difficult to obtain in some cases. Furthermore, the CNNs trained on one specific modality may fail on others for images acquired with different imaging protocols and scanners. Therefore, developing a unified model that can segment the target objects from unpaired multiple modalities is significant for many clinical applications. In this work, we propose a 3D unified generative adversarial network, which unifies the any-to-any modality translation and multimodal segmentation in a single network. Since the anatomical structure is preserved during modality translation, the auxiliary translation task is used to extract the modality-invariant features and generate the additional training data implicitly. To fully utilize the segmentation-related features, we add a cross-task skip connection with feature recalibration from the translation decoder to the segmentation decoder. Experiments on abdominal organ segmentation and brain tumor segmentation indicate that our method outperforms the existing unified methods.     

Fully automatic segmentation of white matter hyperintensities in MR images of the elderly

The role of quantitative image analysis in large clinical trials is continuously increasing. Several methods are available for performing white matter hyperintensity (WMH) volume quantification. They vary in the amount of the human interaction involved. In this paper, we describe a fully automatic segmentation that was used to quantify WMHs in a large clinical trial on elderly subjects. Our segmentation method combines information from 3 different MR images: proton density (PD), T2-weighted and fluid-attenuated inversion recovery (FLAIR) images; our method uses an established artificial intelligent technique (fuzzy inference system) and does not require extensive computations. The reproducibility of the segmentation was evaluated in 9 patients who underwent scan-rescan with repositioning; an inter-class correlation coefficient (ICC) of 0.91 was obtained. The effect of differences in image resolution was tested in 44 patients, scanned with 6- and 3-mm slice thickness FLAIR images; we obtained an ICC value of 0.99. The accuracy of the segmentation was evaluated on 100 patients for whom manual delineation of WMHs was available; the obtained ICC was 0.98 and the similarity index was 0.75. Besides the fact that the approach demonstrated very high volumetric and spatial agreement with expert delineation, the software did not require more than 2 min per patient (from loading the images to saving the results) on a Pentium-4 processor (512 MB RAM).