Multi-atlas segmentation with augmented features for cardiac MR images

Multi-atlas segmentation infers the target image segmentation by combining prior anatomical knowledge encoded in multiple atlases. It has been quite successfully applied to medical image segmentation in the recent years, resulting in highly accurate and robust segmentation for many anatomical structures. However, to guide the label fusion process, most existing multi-atlas segmentation methods only utilise the intensity information within a small patch during the label fusion process and may neglect other useful information such as gradient and contextual information (the appearance of surrounding regions). This paper proposes to combine the intensity, gradient and contextual information into an augmented feature vector and incorporate it into multi-atlas segmentation. Also, it explores the alternative to the K nearest neighbour (KNN) classifier in performing multi-atlas label fusion, by using the support vector machine (SVM) for label fusion instead. Experimental results on a short-axis cardiac MR data set of 83 subjects have demonstrated that the accuracy of multi-atlas segmentation can be significantly improved by using the augmented feature vector. The mean Dice metric of the proposed segmentation framework is 0.81 for the left ventricular myocardium on this data set, compared to 0.79 given by the conventional multi-atlas patch-based segmentation (Coupé et al., 2011; Rousseau et al., 2011). A major contribution of this paper is that it demonstrates that the performance of non-local patch-based segmentation can be improved by using augmented features.     

Groupwise multi-atlas segmentation of the spinal cord’s internal structure

The spinal cord is an essential and vulnerable component of the central nervous system. Differentiating and localizing the spinal cord internal structure (i.e., gray matter vs. white matter) is critical for assessment of therapeutic impacts and determining prognosis of relevant conditions. Fortunately, new magnetic resonance imaging (MRI) sequences enable clinical study of the in vivo spinal cord’s internal structure. Yet, low contrast-to-noise ratio, artifacts, and imaging distortions have limited the applicability of tissue segmentation techniques pioneered elsewhere in the central nervous system. Additionally, due to the inter-subject variability exhibited on cervical MRI, typical deformable volumetric registrations perform poorly, limiting the applicability of a typical multi-atlas segmentation framework. Thus, to date, no automated algorithms have been presented for the spinal cord’s internal structure. Herein, we present a novel slice-based groupwise registration framework for robustly segmenting cervical spinal cord MRI. Specifically, we provide a method for (1) pre-aligning the slice-based atlases into a groupwise-consistent space, (2) constructing a model of spinal cord variability, (3) projecting the target slice into the low-dimensional space using a model-specific registration cost function, and (4) estimating robust segmentation susing geodesically appropriate atlas information. Moreover, the proposed framework provides a natural mechanism for performing atlas selection and initializing the free model parameters in an informed manner. In a cross-validation experiment using 67 MR volumes of the cervical spinal cord, we demonstrate sub-millimetric accuracy, significant quantitative and qualitative improvement over comparable multi-atlas frameworks, and provide insight into the sensitivity of the associated model parameters.     

Multi-slice low-rank tensor decomposition based multi-atlas segmentation: Application to automatic pathological liver CT segmentation

Liver segmentation from abdominal CT images is an essential step for liver cancer computer-aided diagnosis and surgical planning. However, both the accuracy and robustness of existing liver segmentation methods cannot meet the requirements of clinical applications. In particular, for the common clinical cases where the liver tissue contains major pathology, current segmentation methods show poor performance. In this paper, we propose a novel low-rank tensor decomposition (LRTD) based multi-atlas segmentation (MAS) framework that achieves accurate and robust pathological liver segmentation of CT images. Firstly, we propose a multi-slice LRTD scheme to recover the underlying low-rank structure embedded in 3D medical images. It performs the LRTD on small image segments consisting of multiple consecutive image slices. Then, we present an LRTD-based atlas construction method to generate tumor-free liver atlases that mitigates the performance degradation of liver segmentation due to the presence of tumors. Finally, we introduce an LRTD-based MAS algorithm to derive patient-specific liver atlases for each test image, and to achieve accurate pairwise image registration and label propagation. Extensive experiments on three public databases of pathological liver cases validate the effectiveness of the proposed method. Both qualitative and quantitative results demonstrate that, in the presence of major pathology, the proposed method is more accurate and robust than state-of-the-art methods.     

Multi-atlas pancreas segmentation: Atlas selection based on vessel structure

Automated organ segmentation from medical images is an indispensable component for clinical applications such as computer-aided diagnosis (CAD) and computer-assisted surgery (CAS). We utilize a multi-atlas segmentation scheme, which has recently been used in different approaches in the literature to achieve more accurate and robust segmentation of anatomical structures in computed tomography (CT) volume data. Among abdominal organs, the pancreas has large inter-patient variability in its position, size and shape. Moreover, the CT intensity of the pancreas closely resembles adjacent tissues, rendering its segmentation a challenging task. Due to this, conventional intensity-based atlas selection for pancreas segmentation often fails to select atlases that are similar in pancreas position and shape to those of the unlabeled target volume. In this paper, we propose a new atlas selection strategy based on vessel structure around the pancreatic tissue and demonstrate its application to a multi-atlas pancreas segmentation. Our method utilizes vessel structure around the pancreas to select atlases with high pancreatic resemblance to the unlabeled volume. Also, we investigate two types of applications of the vessel structure information to the atlas selection. Our segmentations were evaluated on 150 abdominal contrast-enhanced CT volumes. The experimental results showed that our approach can segment the pancreas with an average Jaccard index of 66.3% and an average Dice overlap coefficient of 78.5%.     

Automatic labeling of cortical sulci using patch- or CNN-based segmentation techniques combined with bottom-up geometric constraints

The extreme variability of the folding pattern of the human cortex makes the recognition of cortical sulci, both automatic and manual, particularly challenging. Reliable identification of the human cortical sulci in its entirety, is extremely difficult and is practiced by only a few experts. Moreover, these sulci correspond to more than a hundred different structures, which makes manual labeling long and fastidious and therefore limits access to large labeled databases to train machine learning. Here, we seek to improve the current model proposed in the Morphologist toolbox, a widely used sulcus recognition toolbox included in the BrainVISA package. Two novel approaches are proposed: patch-based multi-atlas segmentation (MAS) techniques and convolutional neural network (CNN)-based approaches. Both are currently applied for anatomical segmentations because they embed much better representations of inter-subject variability than approaches based on a single template atlas. However, these methods typically focus on voxel-wise labeling, disregarding certain geometrical and topological properties of interest for sulcus morphometry. Therefore, we propose to refine these approaches with domain specific bottom-up geometric constraints provided by the Morphologist toolbox. These constraints are utilized to provide a single sulcus label to each topologically elementary fold, the building blocks of the pattern recognition problem. To eliminate the shortcomings associated with the Morphologist’s pre-segmentation into elementary folds, we complement this regularization scheme using a top-down perspective which triggers an additional cleavage of the elementary folds when required. All the newly proposed models outperform the current Morphologist model, the most efficient being a CNN U-Net-based approach which carries out sulcus recognition within a few seconds.     

Biomechanically constrained non-rigid MR-TRUS prostate registration using deep learning based 3D point cloud matching

A non-rigid MR-TRUS image registration framework is proposed for prostate interventions. The registration framework consists of a convolutional neural networks (CNN) for MR prostate segmentation, a CNN for TRUS prostate segmentation and a point-cloud based network for rapid 3D point cloud matching. Volumetric prostate point clouds were generated from the segmented prostate masks using tetrahedron meshing. The point cloud matching network was trained using deformation field that was generated by finite element analysis. Therefore, the network implicitly models the underlying biomechanical constraint when performing point cloud matching. A total of 50 patients’ datasets were used for the network training and testing. Alignment of prostate shapes after registration was evaluated using three metrics including Dice similarity coefficient (DSC), mean surface distance (MSD) and Hausdorff distance (HD). Internal point-to-point registration accuracy was assessed using target registration error (TRE). Jacobian determinant and strain tensors of the predicted deformation field were calculated to analyze the physical fidelity of the deformation field. On average, the mean and standard deviation were 0.94±0.02, 0.90±0.23 mm, 2.96±1.00 mm and 1.57±0.77 mm for DSC, MSD, HD and TRE, respectively. Robustness of our method to point cloud noise was evaluated by adding different levels of noise to the query point clouds. Our results demonstrated that the proposed method could rapidly perform MR-TRUS image registration with good registration accuracy and robustness.     

High-resolution 3D abdominal segmentation with random patch network fusion

Deep learning for three dimensional (3D) abdominal organ segmentation on high-resolution computed tomography (CT) is a challenging topic, in part due to the limited memory provide by graphics processing units (GPU) and large number of parameters and in 3D fully convolutional networks (FCN). Two prevalent strategies, lower resolution with wider field of view and higher resolution with limited field of view, have been explored but have been presented with varying degrees of success. In this paper, we propose a novel patch-based network with random spatial initialization and statistical fusion on overlapping regions of interest (ROIs). We evaluate the proposed approach using three datasets consisting of 260 subjects with varying numbers of manual labels. Compared with the canonical “coarse-to-fine” baseline methods, the proposed method increases the performance on multi-organ segmentation from 0.799 to 0.856 in terms of mean DSC score (p-value < 0.01 with paired t-test). The effect of different numbers of patches is evaluated by increasing the depth of coverage (expected number of patches evaluated per voxel). In addition, our method outperforms other state-of-the-art methods in abdominal organ segmentation. In conclusion, the approach provides a memory-conservative framework to enable 3D segmentation on high-resolution CT. The approach is compatible with many base network structures, without substantially increasing the complexity during inference.Given a CT scan with at high resolution, a low-res section (left panel) is trained with multi-channel segmentation. The low-res part contains down-sampling and normalization in order to preserve the complete spatial information. Interpolation and random patch sampling (mid panel) is employed to collect patches. The high-dimensional probability maps are acquired (right panel) from integration of all patches on field of views.     

Automatic bladder segmentation from CT images using deep CNN and 3D fully connected CRF-RNN

Purpose

Automatic approach for bladder segmentation from computed tomography (CT) images is highly desirable in clinical practice. It is a challenging task since the bladder usually suffers large variations of appearance and low soft-tissue contrast in CT images. In this study, we present a deep learning-based approach which involves a convolutional neural network (CNN) and a 3D fully connected conditional random fields recurrent neural network (CRF-RNN) to perform accurate bladder segmentation. We also propose a novel preprocessing method, called dual-channel preprocessing, to further advance the segmentation performance of our approach.

Methods

The presented approach works as following: first, we apply our proposed preprocessing method on the input CT image and obtain a dual-channel image which consists of the CT image and an enhanced bladder density map. Second, we exploit a CNN to predict a coarse voxel-wise bladder score map on this dual-channel image. Finally, a 3D fully connected CRF-RNN refines the coarse bladder score map and produce final fine-localized segmentation result.

Results

We compare our approach to the state-of-the-art V-net on a clinical dataset. Results show that our approach achieves superior segmentation accuracy, outperforming the V-net by a significant margin. The Dice Similarity Coefficient of our approach (92.24%) is 8.12% higher than that of the V-net. Moreover, the bladder probability maps performed by our approach present sharper boundaries and more accurate localizations compared with that of the V-net.

Conclusion

Our approach achieves higher segmentation accuracy than the state-of-the-art method on clinical data. Both the dual-channel processing and the 3D fully connected CRF-RNN contribute to this improvement. The united deep network composed of the CNN and 3D CRF-RNN also outperforms a system where the CRF model acts as a post-processing method disconnected from the CNN.

     

Non-local statistical label fusion for multi-atlas segmentation

Multi-atlas segmentation provides a general purpose, fully-automated approach for transferring spatial information from an existing dataset (“atlases”) to a previously unseen context (“target”) through image registration. The method to resolve voxelwise label conflicts between the registered atlases (“label fusion”) has a substantial impact on segmentation quality. Ideally, statistical fusion algorithms (e.g., STAPLE) would result in accurate segmentations as they provide a framework to elegantly integrate models of rater performance. The accuracy of statistical fusion hinges upon accurately modeling the underlying process of how raters err. Despite success on human raters, current approaches inaccurately model multi-atlas behavior as they fail to seamlessly incorporate exogenous intensity information into the estimation process. As a result, locally weighted voting algorithms represent the de facto standard fusion approach in clinical applications. Moreover, regardless of the approach, fusion algorithms are generally dependent upon large atlas sets and highly accurate registration as they implicitly assume that the registered atlases form a collectively unbiased representation of the target. Herein, we propose a novel statistical fusion algorithm, Non-Local STAPLE (NLS). NLS reformulates the STAPLE framework from a non-local means perspective in order to learn what label an atlas would have observed, given perfect correspondence. Through this reformulation, NLS (1) seamlessly integrates intensity into the estimation process, (2) provides a theoretically consistent model of multi-atlas observation error, and (3) largely diminishes the need for large atlas sets and very high-quality registrations. We assess the sensitivity and optimality of the approach and demonstrate significant improvement in two empirical multi-atlas experiments.     

Multi-task learning for segmentation and classification of tumors in 3D automated breast ultrasound images

Tumor classification and segmentation are two important tasks for computer-aided diagnosis (CAD) using 3D automated breast ultrasound (ABUS) images. However, they are challenging due to the significant shape variation of breast tumors and the fuzzy nature of ultrasound images (e.g., low contrast and signal to noise ratio). Considering the correlation between tumor classification and segmentation, we argue that learning these two tasks jointly is able to improve the outcomes of both tasks. In this paper, we propose a novel multi-task learning framework for joint segmentation and classification of tumors in ABUS images. The proposed framework consists of two sub-networks: an encoder-decoder network for segmentation and a light-weight multi-scale network for classification. To account for the fuzzy boundaries of tumors in ABUS images, our framework uses an iterative training strategy to refine feature maps with the help of probability maps obtained from previous iterations. Experimental results based on a clinical dataset of 170 3D ABUS volumes collected from 107 patients indicate that the proposed multi-task framework improves tumor segmentation and classification over the single-task learning counterparts.     

Discriminative confidence estimation for probabilistic multi-atlas label fusion

Quantitative neuroimaging analyses often rely on the accurate segmentation of anatomical brain structures. In contrast to manual segmentation, automatic methods offer reproducible outputs and provide scalability to study large databases. Among existing approaches, multi-atlas segmentation has recently shown to yield state-of-the-art performance in automatic segmentation of brain images. It consists in propagating the labelmaps from a set of atlases to the anatomy of a target image using image registration, and then fusing these multiple warped labelmaps into a consensus segmentation on the target image. Accurately estimating the contribution of each atlas labelmap to the final segmentation is a critical step for the success of multi-atlas segmentation. Common approaches to label fusion either rely on local patch similarity, probabilistic statistical frameworks or a combination of both. In this work, we propose a probabilistic label fusion framework based on atlas label confidences computed at each voxel of the structure of interest. Maximum likelihood atlas confidences are estimated using a supervised approach, explicitly modeling the relationship between local image appearances and segmentation errors produced by each of the atlases. We evaluate different spatial pooling strategies for modeling local segmentation errors. We also present a novel type of label-dependent appearance features based on atlas labelmaps that are used during confidence estimation to increase the accuracy of our label fusion. Our approach is evaluated on the segmentation of seven subcortical brain structures from the MICCAI 2013 SATA Challenge dataset and the hippocampi from the ADNI dataset. Overall, our results indicate that the proposed label fusion framework achieves superior performance to state-of-the-art approaches in the majority of the evaluated brain structures and shows more robustness to registration errors.     

Segmentation of embryonic and fetal 3D ultrasound images based on pixel intensity distributions and shape priors

This paper presents a novel variational segmentation framework combining shape priors and parametric intensity distribution modeling for extracting the fetal envelope on 3D obstetric ultrasound images. To overcome issues related to poor image quality and missing boundaries, we inject three types of information in the segmentation process: tissue-specific parametric modeling of pixel intensities, a shape prior for the fetal envelope and a shape model of the fetus’ back. The shape prior is encoded with Legendre moments and used to constraint the evolution of a level-set function. The back model is used to post-process the segmented fetal envelope. Results are presented on 3D ultrasound data and compared to a set of manual segmentations. The robustness of the algorithm is studied, and both visual and quantitative comparisons show satisfactory results obtained by the proposed method on the tested dataset.     

Automatic abdominal multi-organ segmentation using deep convolutional neural network and time-implicit level sets

Purpose

Multi-organ segmentation from CT images is an essential step for computer-aided diagnosis and surgery planning. However, manual delineation of the organs by radiologists is tedious, time-consuming and poorly reproducible. Therefore, we propose a fully automatic method for the segmentation of multiple organs from three-dimensional abdominal CT images.

Methods

The proposed method employs deep fully convolutional neural networks (CNNs) for organ detection and segmentation, which is further refined by a time-implicit multi-phase evolution method. Firstly, a 3D CNN is trained to automatically localize and delineate the organs of interest with a probability prediction map. The learned probability map provides both subject-specific spatial priors and initialization for subsequent fine segmentation. Then, for the refinement of the multi-organ segmentation, image intensity models, probability priors as well as a disjoint region constraint are incorporated into an unified energy functional. Finally, a novel time-implicit multi-phase level-set algorithm is utilized to efficiently optimize the proposed energy functional model.

Results

Our method has been evaluated on 140 abdominal CT scans for the segmentation of four organs (liver, spleen and both kidneys). With respect to the ground truth, average Dice overlap ratios for the liver, spleen and both kidneys are 96.0, 94.2 and 95.4%, respectively, and average symmetric surface distance is less than 1.3 mm for all the segmented organs. The computation time for a CT volume is 125 s in average. The achieved accuracy compares well to state-of-the-art methods with much higher efficiency.

Conclusion

A fully automatic method for multi-organ segmentation from abdominal CT images was developed and evaluated. The results demonstrated its potential in clinical usage with high effectiveness, robustness and efficiency.

     

Towards cross-modal organ translation and segmentation: A cycle- and shape-consistent generative adversarial network

Synthesized medical images have several important applications. For instance, they can be used as an intermedium in cross-modality image registration or used as augmented training samples to boost the generalization capability of a classifier. In this work, we propose a generic cross-modality synthesis approach with the following targets: 1) synthesizing realistic looking 2D/3D images without needing paired training data, 2) ensuring consistent anatomical structures, which could be changed by geometric distortion in cross-modality synthesis and 3) more importantly, improving volume segmentation by using synthetic data for modalities with limited training samples. We show that these goals can be achieved with an end-to-end 2D/3D convolutional neural network (CNN) composed of mutually-beneficial generators and segmentors for image synthesis and segmentation tasks. The generators are trained with an adversarial loss, a cycle-consistency loss, and also a shape-consistency loss (supervised by segmentors) to reduce the geometric distortion. From the segmentation view, the segmentors are boosted by synthetic data from generators in an online manner. Generators and segmentors prompt each other alternatively in an end-to-end training fashion. We validate our proposed method on three datasets, including cardiovascular CT and magnetic resonance imaging (MRI), abdominal CT and MRI, and mammography X-rays from different data domains, showing both tasks are beneficial to each other and coupling these two tasks results in better performance than solving them exclusively.     

3D image segmentation using combined shape-intensity prior models

Objective Statistical models for medical images have been developed to increase robustness in the segmentation process. In this project, a fully automatic approach to build a statistical shape-intensity model and combine this model with level set segmentation was designed, implemented and tested by applying the algorithm to clinical image data. Methods By using a hierarchical registration approach based on mutual information and demons registration, 3D statistical shape-intensity models were created by applying Principal Component Analysis. Using these models in combination with level set segmentation results in a fully automatic modeling and segmentation pipeline. Results Examples for shape-intensity models were synthesized and these models were used to automatically segment 3D MRI and CT images. Quantitative evaluation of the framework was performed by comparing automatic segmentation results to segmentation results of medical experts. Conclusion Evaluation tests in which this method was used for the automatic segmentation of femora and cardiac MRI endocardial surfaces are very promising. The implementation of an additional cost function term and the addition of information about the surroundings of an organ in the model are currently under development.     

Atlas-ISTN: Joint segmentation,registration and atlas construction with image-and-spatial transformer networks

Deep learning models for semantic segmentation are able to learn powerful representations for pixel-wise predictions, but are sensitive to noise at test time and may lead to implausible topologies. Image registration models on the other hand are able to warp known topologies to target images as a means of segmentation, but typically require large amounts of training data, and have not widely been benchmarked against pixel-wise segmentation models. We propose the Atlas Image-and-Spatial Transformer Network (Atlas-ISTN), a framework that jointly learns segmentation and registration on 2D and 3D image data, and constructs a population-derived atlas in the process. Atlas-ISTN learns to segment multiple structures of interest and to register the constructed atlas labelmap to an intermediate pixel-wise segmentation. Additionally, Atlas-ISTN allows for test time refinement of the model’s parameters to optimize the alignment of the atlas labelmap to an intermediate pixel-wise segmentation. This process both mitigates for noise in the target image that can result in spurious pixel-wise predictions, as well as improves upon the one-pass prediction of the model. Benefits of the Atlas-ISTN framework are demonstrated qualitatively and quantitatively on 2D synthetic data and 3D cardiac computed tomography and brain magnetic resonance image data, out-performing both segmentation and registration baseline models. Atlas-ISTN also provides inter-subject correspondence of the structures of interest.     

MIDeepSeg: Minimally interactive segmentation of unseen objects from medical images using deep learning

Segmentation of organs or lesions from medical images plays an essential role in many clinical applications such as diagnosis and treatment planning. Though Convolutional Neural Networks (CNN) have achieved the state-of-the-art performance for automatic segmentation, they are often limited by the lack of clinically acceptable accuracy and robustness in complex cases. Therefore, interactive segmentation is a practical alternative to these methods. However, traditional interactive segmentation methods require a large number of user interactions, and recently proposed CNN-based interactive segmentation methods are limited by poor performance on previously unseen objects. To solve these problems, we propose a novel deep learning-based interactive segmentation method that not only has high efficiency due to only requiring clicks as user inputs but also generalizes well to a range of previously unseen objects. Specifically, we first encode user-provided interior margin points via our proposed exponentialized geodesic distance that enables a CNN to achieve a good initial segmentation result of both previously seen and unseen objects, then we use a novel information fusion method that combines the initial segmentation with only a few additional user clicks to efficiently obtain a refined segmentation. We validated our proposed framework through extensive experiments on 2D and 3D medical image segmentation tasks with a wide range of previously unseen objects that were not present in the training set. Experimental results showed that our proposed framework 1) achieves accurate results with fewer user interactions and less time compared with state-of-the-art interactive frameworks and 2) generalizes well to previously unseen objects.     

Volume segmentation and reconstruction from freehand three-dimensional ultrasound data with application to ovarian follicle measurement

This article presents a semi-automatic method for segmentation and reconstruction of freehand three-dimensional (3D) ultrasound data. The method incorporates a number of interesting features within the level-set framework: First, segmentation is carried out using region competition, requiring multiple distinct and competing regions to be encoded within the framework. This region competition uses a simple dot-product based similarity measure to compare intensities within each region. In addition, segmentation and surface reconstruction is performed within the 3D domain to take advantage of the additional spatial information available. This means that the method must interpolate the surface where there are gaps in the data, a feature common to freehand 3D ultrasound reconstruction. Finally, although the level-set method is restricted to a voxel grid, no assumption is made that the data being segmented will conform to this grid and may be segmented in its world-reference position. The volume reconstruction method is demonstrated in vivo for the volume measurement of ovarian follicles. The 3D reconstructions produce a lower error variance than the current clinical measurement based on a mean diameter estimated from two-dimensional (2D) images. However, both the clinical measurement and the semi-automatic method appear to underestimate the true follicular volume.     

Active deformation fields: dense deformation field estimation for atlas-based segmentation using the active contour framework

This paper presents a new and original variational framework for atlas-based segmentation. The proposed framework integrates both the active contour framework, and the dense deformation fields of optical flow framework. This framework is quite general and encompasses many of the state-of-the-art atlas-based segmentation methods. It also allows to perform the registration of atlas and target images based on only selected structures of interest. The versatility and potentiality of the proposed framework are demonstrated by presenting three diverse applications: In the first application, we show how the proposed framework can be used to simulate the growth of inconsistent structures like a tumor in an atlas. In the second application, we estimate the position of nonvisible brain structures based on the surrounding structures and validate the results by comparing with other methods. In the final application, we present the segmentation of lymph nodes in the Head and Neck CT images, and demonstrate how multiple registration forces can be used in this framework in an hierarchical manner.