FULLY CONVOLUTIONAL NETWORKS FOR MULTI-MODALITY ISOINTENSE INFANT BRAIN IMAGE SEGMENTATION

MhURI:A Supervised Segmentation Approach to Leverage Salient Brain Tissues in Magnetic Resonance Images

BACKGROUND AND OBJECTIVES

Accurate segmentation of critical tissues from a brain MRI is pivotal for characterization and quantitative pattern analysis of the human brain and thereby, identifies the earliest signs of various neurodegenerative diseases. To date, in most cases, it is done manually by the radiologists. The overwhelming workload in some of the thickly populated nations may cause exhaustion leading to interruption for the doctors, which may pose a continuing threat to patient safety. A novel fusion method called U-Net inception based on 3D convolutions and transition layers is proposed to address this issue.

METHODS

A 3D deep learning method called Multi headed U-Net with Residual Inception (MhURI) accompanied by Morphological Gradient channel for brain tissue segmentation is proposed, which incorporates Residual Inception 2-Residual (RI2R) module as the basic building block. The model exploits the benefits of morphological pre-processing for structural enhancement of MR images. A multi-path data encoding pipeline is introduced on top of the U-Net backbone, which encapsulates initial global features and captures the information from each MRI modality.

RESULTS

The proposed model has accomplished encouraging outcomes, which appreciates the adequacy in terms of some of the established quality metrices when compared with some of the state-of-the-art methods while evaluating with respect to two popular publicly available data sets.

CONCLUSION

The model is entirely automatic and able to segment gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) from brain MRI effectively with sufficient accuracy. Hence, it may be considered to be a potential computer-aided diagnostic (CAD) tool for radiologists and other medical practitioners in their clinical diagnosis workflow.

Analysis of the nonperfused volume ratio of adenomyosis from MRI images based on fewshot learning

The nonperfused volume (NPV) ratio is the key to the success of high intensity focused ultrasound (HIFU) ablation treatment of adenomyosis. However, there are no qualitative interpretation standards for predicting the NPV ratio of adenomyosis using magnetic resonance imaging (MRI) before HIFU ablation treatment, which leading to inter-reader variability. Convolutional neural networks (CNNs) have achieved state-of-the-art performance in the automatic disease diagnosis of MRI. Since the use of HIFU to treat adenomyosis is a novel treatment, there is not enough MRI data to support CNNs. We proposed a novel few-shot learning framework that extends CNNs to predict NPV ratio of HIFU ablation treatment for adenomyosis. We collected a dataset from 208 patients with adenomyosis who underwent MRI examination before and after HIFU treatment. Our proposed method was trained and evaluated by fourfold cross validation. This framework obtained sensitivity of 85.6%, 89.6% and 92.8% at 0.799, 0.980 and 1.180 FPs per patient. In the receiver operating characteristics analysis for NPV ratio of adenomyosis, our proposed method received the area under the curve of 0.8233, 0.8289, 0.8412, 0.8319, 0.7010, 0.7637, 0.8375, 0.8219, 0.8207, 0.9812 for the classifications of NPV ratio interval [0%-10%), [10%-20%), …, [90%-100%], respectively. The present study demonstrated that few-shot learning on NPV ratio prediction of HIFU ablation treatment for adenomyosis may contribute to the selection of eligible patients and the pre-judgment of clinical efficacy.

Segmentation of cellular patterns in confocal images of melanocytic lesions in vivo via a multiscale encoder-decoder network (MED-Net)

In-vivo optical microscopy is advancing into routine clinical practice for non-invasively guiding diagnosis and treatment of cancer and other diseases, and thus beginning to reduce the need for traditional biopsy. However, reading and analysis of the optical microscopic images are generally still qualitative, relying mainly on visual examination. Here we present an automated semantic segmentation method called "Multiscale Encoder-Decoder Network (MED-Net)" that provides pixel-wise labeling into classes of patterns in a quantitative manner. The novelty in our approach is the modeling of textural patterns at multiple scales (magnifications, resolutions). This mimics the traditional procedure for examining pathology images, which routinely starts with low magnification (low resolution, large field of view) followed by closer inspection of suspicious areas with higher magnification (higher resolution, smaller fields of view). We trained and tested our model on non-overlapping partitions of 117 reflectance confocal microscopy (RCM) mosaics of melanocytic lesions, an extensive dataset for this application, collected at four clinics in the US, and two in Italy. With patient-wise cross-validation, we achieved pixel-wise mean sensitivity and specificity of 74% and 92%, respectively, with 0.74 Dice coefficient over six classes. In the scenario, we partitioned the data clinic-wise and tested the generalizability of the model over multiple clinics. In this setting, we achieved pixel-wise mean sensitivity and specificity of 77% and 94%, respectively, with 0.77 Dice coefficient. We compared MED-Net against the state-of-the-art semantic segmentation models and achieved better quantitative segmentation performance. Our results also suggest that, due to its nested multiscale architecture, the MED-Net model annotated RCM mosaics more coherently, avoiding unrealistic-fragmented annotations.

Deep HT: A deep neural network for diagnose on MR images of tumors of the hand

BACKGROUND

There are many types of hand tumors, and it is often difficult for imaging diagnosticians to make a correct diagnosis, which can easily lead to misdiagnosis and delay in treatment. Thus in this paper, we propose a deep neural network for diagnose on MR Images of tumors of the hand in order to better define preoperative diagnosis and standardize surgical treatment.

METHODS

We collected MRI figures of 221 patients with hand tumors from one medical center from 2016 to 2019, invited medical experts to annotate the images to form the annotation data set. Then the original image is preprocessed to get the image data set. The data set is randomly divided into ten parts, nine for training and one for test. Next, the data set is input into the neural network system for testing. Finally, average the results of ten experiments as an estimate of the accuracy of the algorithm.

RESULTS

This research uses 221 images as dataset and the system shows an average confidence level of 71.6% in segmentation of hand tumors. The segmented tumor regions are validated through ground truth analysis and manual analysis by a radiologist.

CONCLUSIONS

With the recent advances in convolutional neural networks, vast improvements have been made for image segmentation, mainly based on the skip-connection-linked encoder decoder deep architectures. Therefore, in this paper, we propose an automatic segmentation method based on DeepLab v3+ and achieved a good diagnostic accuracy rate.

Variability and reproducibility in deep learning for medical image segmentation

Medical image segmentation is an important tool for current clinical applications. It is the backbone of numerous clinical diagnosis methods, oncological treatments and computer-integrated surgeries. A new class of machine learning algorithm, deep learning algorithms, outperforms the results of classical segmentation in terms of accuracy. However, these techniques are complex and can have a high range of variability, calling the reproducibility of the results into question. In this article, through a literature review, we propose an original overview of the sources of variability to better understand the challenges and issues of reproducibility related to deep learning for medical image segmentation. Finally, we propose 3 main recommendations to address these potential issues: (1) an adequate description of the framework of deep learning, (2) a suitable analysis of the different sources of variability in the framework of deep learning, and (3) an efficient system for evaluating the segmentation results.

Automatic segmentation of brain MRI using a novel patch-wise U-net deep architecture

Accurate segmentation of brain magnetic resonance imaging (MRI) is an essential step in quantifying the changes in brain structure. Deep learning in recent years has been extensively used for brain image segmentation with highly promising performance. In particular, the U-net architecture has been widely used for segmentation in various biomedical related fields. In this paper, we propose a patch-wise U-net architecture for the automatic segmentation of brain structures in structural MRI. In the proposed brain segmentation method, the non-overlapping patch-wise U-net is used to overcome the drawbacks of conventional U-net with more retention of local information. In our proposed method, the slices from an MRI scan are divided into non-overlapping patches that are fed into the U-net model along with their corresponding patches of ground truth so as to train the network. The experimental results show that the proposed patch-wise U-net model achieves a Dice similarity coefficient (DSC) score of 0.93 in average and outperforms the conventional U-net and the SegNet-based methods by 3% and 10%, respectively, for on Open Access Series of Imaging Studies (OASIS) and Internet Brain Segmentation Repository (IBSR) dataset.

Semantic Segmentation of White Matter in FDG-PET Using Generative Adversarial Network

In the diagnosis of neurodegenerative disorders, F-18 fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) is used for its ability to detect functional changes at early stages of disease process. However, anatomical information from another modality (CT or MRI) is still needed to properly interpret and localize the radiotracer uptake due to its low spatial resolution. Lack of structural information limits segmentation and accurate quantification of the 18F-FDG PET/CT. The correct segmentation of the brain compartment in 18F-FDG PET/CT will enable the quantitative analysis of the 18F-FDG PET/CT scan alone. In this paper, we propose a method to segment white matter in 18F-FDG PET/CT images using generative adversarial network (GAN). The segmentation result of GAN model was evaluated using evaluation parameters such as dice, AUC-PR, precision, and recall. It was also compared with other deep learning methods. As a result, the proposed method achieves superior segmentation accuracy and reliability compared with other deep learning methods.

Unpaired Multi-Modal Segmentation via Knowledge Distillation

Multi-modal learning is typically performed with network architectures containing modality-specific layers and shared layers, utilizing co-registered images of different modalities. We propose a novel learning scheme for unpaired cross-modality image segmentation, with a highly compact architecture achieving superior segmentation accuracy. In our method, we heavily reuse network parameters, by sharing all convolutional kernels across CT and MRI, and only employ modality-specific internal normalization layers which compute respective statistics. To effectively train such a highly compact model, we introduce a novel loss term inspired by knowledge distillation, by explicitly constraining the KL-divergence of our derived prediction distributions between modalities. We have extensively validated our approach on two multi-class segmentation problems: i) cardiac structure segmentation, and ii) abdominal organ segmentation. Different network settings, i.e., 2D dilated network and 3D U-net, are utilized to investigate our method's general efficacy. Experimental results on both tasks demonstrate that our novel multi-modal learning scheme consistently outperforms single-modal training and previous multi-modal approaches.

MRI Segmentation and Classification of Human Brain Using Deep Learning for Diagnosis of Alzheimer's Disease: A Survey

Many neurological diseases and delineating pathological regions have been analyzed, and the anatomical structure of the brain researched with the aid of magnetic resonance imaging (MRI). It is important to identify patients with Alzheimer's disease (AD) early so that preventative measures can be taken. A detailed analysis of the tissue structures from segmented MRI leads to a more accurate classification of specific brain disorders. Several segmentation methods to diagnose AD have been proposed with varying complexity. Segmentation of the brain structure and classification of AD using deep learning approaches has gained attention as it can provide effective results over a large set of data. Hence, deep learning methods are now preferred over state-of-the-art machine learning methods. We aim to provide an outline of current deep learning-based segmentation approaches for the quantitative analysis of brain MRI for the diagnosis of AD. Here, we report how convolutional neural network architectures are used to analyze the anatomical brain structure and diagnose AD, discuss how brain MRI segmentation improves AD classification, describe the state-of-the-art approaches, and summarize their results using publicly available datasets. Finally, we provide insight into current issues and discuss possible future research directions in building a computer-aided diagnostic system for AD.

Machine learning techniques for biomedical image segmentation: An overview of technical aspects and introduction to state-of-art applications

In recent years, significant progress has been made in developing more accurate and efficient machine learning algorithms for segmentation of medical and natural images. In this review article, we highlight the imperative role of machine learning algorithms in enabling efficient and accurate segmentation in the field of medical imaging. We specifically focus on several key studies pertaining to the application of machine learning methods to biomedical image segmentation. We review classical machine learning algorithms such as Markov random fields, k-means clustering, random forest, etc. Although such classical learning models are often less accurate compared to the deep-learning techniques, they are often more sample efficient and have a less complex structure. We also review different deep-learning architectures, such as the artificial neural networks (ANNs), the convolutional neural networks (CNNs), and the recurrent neural networks (RNNs), and present the segmentation results attained by those learning models that were published in the past 3 yr. We highlight the successes and limitations of each machine learning paradigm. In addition, we discuss several challenges related to the training of different machine learning models, and we present some heuristics to address those challenges.

Neural Segmentation of Seeding ROIs (sROIs) for Pre-Surgical Brain Tractography

White matter tractography mapping is an important tool for neuro-surgical planning and navigation. It relies on the accurate manual delineation of anatomical seeding ROIs (sROIs) by neuroanatomy experts. Stringent pre-operative time-constraints and limited availability of experts suggest that automation tools are strongly needed for the task. In this article, we propose and compare several multi-modal fully convolutional network architectures for segmentation of sROIs. Inspired by their manual segmentation practice, anatomical information from T1w maps is fused by the network with directionally encoded color (DEC) maps to compute the segmentation. Qualitative and quantitative validation was performed on image data from 75 real tumor resection candidates for the sROIs of the motor tract, the arcuate fasciculus, and optic radiation. Favorable comparison was also obtained with state-of-the-art methods for the tumor dataset as well as the ISMRM 2017 traCED challenge dataset. The proposed networks showed promising results, indicating they may significantly improve the efficiency of pre-surgical tractography mapping, without compromising its quality.

A 3D Spatially Weighted Network for Segmentation of Brain Tissue From MRI

The segmentation of brain tissue in MRI is valuable for extracting brain structure to aid diagnosis, treatment and tracking the progression of different neurologic diseases. Medical image data are volumetric and some neural network models for medical image segmentation have addressed this using a 3D convolutional architecture. However, this volumetric spatial information has not been fully exploited to enhance the representative ability of deep networks, and these networks have not fully addressed the practical issues facing the analysis of multimodal MRI data. In this paper, we propose a spatially-weighted 3D network (SW-3D-UNet) for brain tissue segmentation of single-modality MRI, and extend it using multimodality MRI data. We validate our model on the MRBrainS13 and MALC12 datasets. This unpublished model ranked first on the leaderboard of the MRBrainS13 Challenge.

Role of deep learning in infant brain MRI analysis

Deep learning algorithms and in particular convolutional networks have shown tremendous success in medical image analysis applications, though relatively few methods have been applied to infant MRI data due numerous inherent challenges such as inhomogenous tissue appearance across the image, considerable image intensity variability across the first year of life, and a low signal to noise setting. This paper presents methods addressing these challenges in two selected applications, specifically infant brain tissue segmentation at the isointense stage and presymptomatic disease prediction in neurodevelopmental disorders. Corresponding methods are reviewed and compared, and open issues are identified, namely low data size restrictions, class imbalance problems, and lack of interpretation of the resulting deep learning solutions. We discuss how existing solutions can be adapted to approach these issues as well as how generative models seem to be a particularly strong contender to address them.

Deep Learning Techniques for Medical Image Segmentation: Achievements and Challenges

Deep learning-based image segmentation is by now firmly established as a robust tool in image segmentation. It has been widely used to separate homogeneous areas as the first and critical component of diagnosis and treatment pipeline. In this article, we present a critical appraisal of popular methods that have employed deep-learning techniques for medical image segmentation. Moreover, we summarize the most common challenges incurred and suggest possible solutions.

Accurate colorectal tumor segmentation for CT scans based on the label assignment generative adversarial network

PURPOSE

Colorectal tumor segmentation is an important step in the analysis and diagnosis of colorectal cancer. This task is a time consuming one since it is often performed manually by radiologists. This paper presents an automatic postprocessing module to refine the segmentation of deep networks. The label assignment generative adversarial network (LAGAN) is improved from the generative adversarial network (GAN) and assigns labels to the outputs of deep networks. We apply the LAGAN to segment colorectal tumors in computed tomography (CT) scans and explore the performances of different combinations of deep networks.

MATERIAL AND METHODS

A total of 223 patients with colorectal cancer (CRC) are enrolled in the study. The CT scans of the colorectal tumors are first segmented by FCN32 and Unet separately, which output probabilistic maps. Then, the probabilistic maps are labeled by the LAGAN and finally, the binary segmentation results are obtained. The LAGAN consists of a generating model and a discriminating model. The generating model utilizes the probabilistic maps from deep networks to imitate the distribution of the ground truths, and the discriminating model attempts to distinguish generations and ground truths. Through competitive training, the generating model of the LAGAN can realize label assignments for the probabilistic maps.

RESULTS

The LAGAN increases the DSC of FCN32 from 81.83% ± 0.35% to 90.82% ± 0.36%. In the Unet-based segmentation, the LAGAN increases the DSC from 86.67% ± 0.70% to 91.54% ± 0.53%. It takes approximately 10 ms to refine a single CT slice.

CONCLUSIONS

The results demonstrate that the LAGAN is a robust and flexible module, which can be used to refine the segmentation of diverse deep networks. Compared with other networks, the LAGAN can achieve desirable segmented accuracy for colorectal tumors.

High-Resolution Encoder-Decoder Networks for Low-Contrast Medical Image Segmentation

Automatic image segmentation is an essential step for many medical image analysis applications, include computer-aided radiation therapy, disease diagnosis, and treatment effect evaluation. One of the major challenges for this task is the blurry nature of medical images (e.g., CT, MR and, microscopic images), which can often result in low-contrast and vanishing boundaries. With the recent advances in convolutional neural networks, vast improvements have been made for image segmentation, mainly based on the skip-connection-linked encoder-decoder deep architectures. However, in many applications (with adjacent targets in blurry images), these models often fail to accurately locate complex boundaries and properly segment tiny isolated parts. In this paper, we aim to provide a method for blurry medical image segmentation and argue that skip connections are not enough to help accurately locate indistinct boundaries. Accordingly, we propose a novel high-resolution multi-scale encoder-decoder network (HMEDN), in which multi-scale dense connections are introduced for the encoder-decoder structure to finely exploit comprehensive semantic information. Besides skip connections, extra deeply-supervised high-resolution pathways (comprised of densely connected dilated convolutions) are integrated to collect high-resolution semantic information for accurate boundary localization. These pathways are paired with a difficulty-guided cross-entropy loss function and a contour regression task to enhance the quality of boundary detection. Extensive experiments on a pelvic CT image dataset, a multi-modal brain tumor dataset, and a cell segmentation dataset show the effectiveness of our method for 2D/3D semantic segmentation and 2D instance segmentation, respectively. Our experimental results also show that besides increasing the network complexity, raising the resolution of semantic feature maps can largely affect the overall model performance. For different tasks, finding a balance between these two factors can further improve the performance of the corresponding network.

3D Auto-Context-Based Locality Adaptive Multi-Modality GANs for PET Synthesis

Positron emission tomography (PET) has been substantially used recently. To minimize the potential health risk caused by the tracer radiation inherent to PET scans, it is of great interest to synthesize the high-quality PET image from the low-dose one to reduce the radiation exposure. In this paper, we propose a 3D auto-context-based locality adaptive multi-modality generative adversarial networks model (LA-GANs) to synthesize the high-quality FDG PET image from the low-dose one with the accompanying MRI images that provide anatomical information. Our work has four contributions. First, different from the traditional methods that treat each image modality as an input channel and apply the same kernel to convolve the whole image, we argue that the contributions of different modalities could vary at different image locations, and therefore a unified kernel for a whole image is not optimal. To address this issue, we propose a locality adaptive strategy for multi-modality fusion. Second, we utilize 1 ×1 ×1 kernel to learn this locality adaptive fusion so that the number of additional parameters incurred by our method is kept minimum. Third, the proposed locality adaptive fusion mechanism is learned jointly with the PET image synthesis in a 3D conditional GANs model, which generates high-quality PET images by employing large-sized image patches and hierarchical features. Fourth, we apply the auto-context strategy to our scheme and propose an auto-context LA-GANs model to further refine the quality of synthesized images. Experimental results show that our method outperforms the traditional multi-modality fusion methods used in deep networks, as well as the state-of-the-art PET estimation approaches.

Multi-atlas active contour segmentation method using template optimization algorithm

Background

Brain image segmentation is the basis and key to brain disease diagnosis, treatment planning and tissue 3D reconstruction. The accuracy of segmentation directly affects the therapeutic effect. Manual segmentation of these images is time-consuming and subjective. Therefore, it is important to research semi-automatic and automatic image segmentation methods. In this paper, we propose a semi-automatic image segmentation method combined with a multi-atlas registration method and an active contour model (ACM).

Method

We propose a multi-atlas active contour segmentation method using a template optimization algorithm. First, a multi-atlas registration method is used to obtain the prior shape information of the target tissue, and then a label fusion algorithm is used to generate the initial template. Second, a template optimization algorithm is used to reduce the multi-atlas registration errors and generate the initial active contour (IAC). Finally, a ACM is used to segment the target tissue.

Results

The proposed method was applied to the challenging publicly available MR datasets IBSR and MRBrainS13. In the MRBrainS13 datasets, we obtained an average thalamus Dice similarity coefficient of 0.927 ± 0.014 and an average Hausdorff distance (HD) of 2.92 ± 0.53. In the IBSR datasets, we obtained a white matter (WM) average Dice similarity coefficient of 0.827 ± 0.04 and a gray gray matter (GM) average Dice similarity coefficient of 0.853 ± 0.03.

Conclusion

In this paper, we propose a semi-automatic brain image segmentation method. The main contributions of this paper are as follows: 1) Our method uses a multi-atlas registration method based on affine transformation, which effectively reduces the multi-atlas registration time compared to the complex nonlinear registration method. The average registration time of each target image in the IBSR datasets is 255 s, and the average registration time of each target image in the MRBrainS13 datasets is 409 s. 2) We used a template optimization algorithm to improve registration error and generate a continuous IAC. 3) Finally, we used a ACM to segment the target tissue and obtain a smooth continuous target contour.

HyperDense-Net: A Hyper-Densely Connected CNN for Multi-Modal Image Segmentation

Recently, dense connections have attracted substantial attention in computer vision because they facilitate gradient flow and implicit deep supervision during training. Particularly, DenseNet that connects each layer to every other layer in a feed-forward fashion and has shown impressive performances in natural image classification tasks. We propose HyperDenseNet, a 3-D fully convolutional neural network that extends the definition of dense connectivity to multi-modal segmentation problems. Each imaging modality has a path, and dense connections occur not only between the pairs of layers within the same path but also between those across different paths. This contrasts with the existing multi-modal CNN approaches, in which modeling several modalities relies entirely on a single joint layer (or level of abstraction) for fusion, typically either at the input or at the output of the network. Therefore, the proposed network has total freedom to learn more complex combinations between the modalities, within and in-between all the levels of abstraction, which increases significantly the learning representation. We report extensive evaluations over two different and highly competitive multi-modal brain tissue segmentation challenges, iSEG 2017 and MRBrainS 2013, with the former focusing on six month infant data and the latter on adult images. HyperDenseNet yielded significant improvements over many state-of-the-art segmentation networks, ranking at the top on both benchmarks. We further provide a comprehensive experimental analysis of features re-use, which confirms the importance of hyper-dense connections in multi-modal representation learning. Our code is publicly available.

STRAINet: Spatially Varying sTochastic Residual AdversarIal Networks for MRI Pelvic Organ Segmentation

Accurate segmentation of pelvic organs is important for prostate radiation therapy. Modern radiation therapy starts to use a magnetic resonance image (MRI) as an alternative to computed tomography image because of its superior soft tissue contrast and also free of risk from radiation exposure. However, segmentation of pelvic organs from MRI is a challenging problem due to inconsistent organ appearance across patients and also large intrapatient anatomical variations across treatment days. To address such challenges, we propose a novel deep network architecture, called "Spatially varying sTochastic Residual AdversarIal Network" (STRAINet), to delineate pelvic organs from MRI in an end-to-end fashion. Compared to the traditional fully convolutional networks (FCN), the proposed architecture has two main contributions: 1) inspired by the recent success of residual learning, we propose an evolutionary version of the residual unit, i.e., stochastic residual unit, and use it to the plain convolutional layers in the FCN. We further propose long-range stochastic residual connections to pass features from shallow layers to deep layers; and 2) we propose to integrate three previously proposed network strategies to form a new network for better medical image segmentation: a) we apply dilated convolution in the smallest resolution feature maps, so that we can gain a larger receptive field without overly losing spatial information; b) we propose a spatially varying convolutional layer that adapts convolutional filters to different regions of interest; and c) an adversarial network is proposed to further correct the segmented organ structures. Finally, STRAINet is used to iteratively refine the segmentation probability maps in an autocontext manner. Experimental results show that our STRAINet achieved the state-of-the-art segmentation accuracy. Further analysis also indicates that our proposed network components contribute most to the performance.

3-D Fully Convolutional Networks for Multimodal Isointense Infant Brain Image Segmentation

Accurate segmentation of infant brain images into different regions of interest is one of the most important fundamental steps in studying early brain development. In the isointense phase (approximately 6-8 months of age), white matter and gray matter exhibit similar levels of intensities in magnetic resonance (MR) images, due to the ongoing myelination and maturation. This results in extremely low tissue contrast and thus makes tissue segmentation very challenging. Existing methods for tissue segmentation in this isointense phase usually employ patch-based sparse labeling on single modality. To address the challenge, we propose a novel 3-D multimodal fully convolutional network (FCN) architecture for segmentation of isointense phase brain MR images. Specifically, we extend the conventional FCN architectures from 2-D to 3-D, and, rather than directly using FCN, we intuitively integrate coarse (naturally high-resolution) and dense (highly semantic) feature maps to better model tiny tissue regions, in addition, we further propose a transformation module to better connect the aggregating layers; we also propose a fusion module to better serve the fusion of feature maps. We compare the performance of our approach with several baseline and state-of-the-art methods on two sets of isointense phase brain images. The comparison results show that our proposed 3-D multimodal FCN model outperforms all previous methods by a large margin in terms of segmentation accuracy. In addition, the proposed framework also achieves faster segmentation results compared to all other methods. Our experiments further demonstrate that: 1) carefully integrating coarse and dense feature maps can considerably improve the segmentation performance; 2) batch normalization can speed up the convergence of the networks, especially when hierarchical feature aggregations occur; and 3) integrating multimodal information can further boost the segmentation performance.

Pelvic Organ Segmentation Using Distinctive Curve Guided Fully Convolutional Networks

Accurate segmentation of pelvic organs (i.e., prostate, bladder, and rectum) from CT image is crucial for effective prostate cancer radiotherapy. However, it is a challenging task due to: 1) low soft tissue contrast in CT images and 2) large shape and appearance variations of pelvic organs. In this paper, we employ a two-stage deep learning-based method, with a novel distinctive curve-guided fully convolutional network (FCN), to solve the aforementioned challenges. Specifically, the first stage is for fast and robust organ detection in the raw CT images. It is designed as a coarse segmentation network to provide region proposals for three pelvic organs. The second stage is for fine segmentation of each organ, based on the region proposal results. To better identify those indistinguishable pelvic organ boundaries, a novel morphological representation, namely, distinctive curve, is also introduced to help better conduct the precise segmentation. To implement this, in this second stage, a multi-task FCN is initially utilized to learn the distinctive curve and the segmentation map separately and then combine these two tasks to produce accurate segmentation map. The final segmentation results of all three pelvic organs are generated by a weighted max-voting strategy. We have conducted exhaustive experiments on a large and diverse pelvic CT data set for evaluating our proposed method. The experimental results demonstrate that our proposed method is accurate and robust for this challenging segmentation task, by also outperforming the state-of-the-art segmentation methods.

Nasopharyngeal carcinoma segmentation based on enhanced convolutional neural networks using multi-modal metric learning

Multi-modality examinations have been extensively applied in current clinical cancer management. Leveraging multi-modality medical images can be highly beneficial for automated tumor segmentation as they provide complementary information that could make the segmentation of tumors more accurate. This paper investigates CNN-based methods for automated nasopharyngeal carcinoma (NPC) segmentation using computed tomography (CT) and magnetic resonance (MR) images. Specially, a multi-modality convolutional neural network (M-CNN) is designed to jointly learn a multi-modal similarity metric and segmentation of paired CT-MR images. By jointly optimizing the similarity learning error and the segmentation error, the feature learning processes of both modalities are mutually guided. In doing so, the segmentation sub-networks are able to take advantage of the other modality's information. Considering that each modality possesses certain distinctive characteristics, we combine the higher-layer features extracted by a single-modality CNN (S-CNN) and M-CNN to form a combined CNN (C-CNN) for each modality, which is able to further utilize the complementary information of different modalities and improve the segmentation performance. The proposed M-CNN and C-CNN were evaluated on 90 CT-MR images of NPC patients. Experimental results demonstrate that our methods achieve improved segmentation performance compared to their counterparts without multi-modal information fusion and the existing CNN-based multi-modality segmentation methods.

Multiorgan segmentation using distance-aware adversarial networks

Segmentation of organs at risk (OAR) in computed tomography (CT) is of vital importance in radiotherapy treatment. This task is time consuming and for some organs, it is very challenging due to low-intensity contrast in CT. We propose a framework to perform the automatic segmentation of multiple OAR: esophagus, heart, trachea, and aorta. Different from previous works using deep learning techniques, we make use of global localization information, based on an original distance map that yields not only the localization of each organ, but also the spatial relationship between them. Instead of segmenting directly the organs, we first generate the localization map by minimizing a reconstruction error within an adversarial framework. This map that includes localization information of all organs is then used to guide the segmentation task in a fully convolutional setting. Experimental results show encouraging performance on CT scans of 60 patients totaling 11,084 slices in comparison with other state-of-the-art methods.

Medical Image Synthesis with Deep Convolutional Adversarial Networks

Medical imaging plays a critical role in various clinical applications. However, due to multiple considerations such as cost and radiation dose, the acquisition of certain image modalities may be limited. Thus, medical image synthesis can be of great benefit by estimating a desired imaging modality without incurring an actual scan. In this paper, we propose a generative adversarial approach to address this challenging problem. Specifically, we train a fully convolutional network (FCN) to generate a target image given a source image. To better model a nonlinear mapping from source to target and to produce more realistic target images, we propose to use the adversarial learning strategy to better model the FCN. Moreover, the FCN is designed to incorporate an image-gradient-difference-based loss function to avoid generating blurry target images. Long-term residual unit is also explored to help the training of the network. We further apply Auto-Context Model to implement a context-aware deep convolutional adversarial network. Experimental results show that our method is accurate and robust for synthesizing target images from the corresponding source images. In particular, we evaluate our method on three datasets, to address the tasks of generating CT from MRI and generating 7T MRI from 3T MRI images. Our method outperforms the state-of-the-art methods under comparison in all datasets and tasks.

Convolutional Neural Networks-Based MRI Image Analysis for the Alzheimer's Disease Prediction From Mild Cognitive Impairment

Mild cognitive impairment (MCI) is the prodromal stage of Alzheimer's disease (AD). Identifying MCI subjects who are at high risk of converting to AD is crucial for effective treatments. In this study, a deep learning approach based on convolutional neural networks (CNN), is designed to accurately predict MCI-to-AD conversion with magnetic resonance imaging (MRI) data. First, MRI images are prepared with age-correction and other processing. Second, local patches, which are assembled into 2.5 dimensions, are extracted from these images. Then, the patches from AD and normal controls (NC) are used to train a CNN to identify deep learning features of MCI subjects. After that, structural brain image features are mined with FreeSurfer to assist CNN. Finally, both types of features are fed into an extreme learning machine classifier to predict the AD conversion. The proposed approach is validated on the standardized MRI datasets from the Alzheimer's Disease Neuroimaging Initiative (ADNI) project. This approach achieves an accuracy of 79.9% and an area under the receiver operating characteristic curve (AUC) of 86.1% in leave-one-out cross validations. Compared with other state-of-the-art methods, the proposed one outperforms others with higher accuracy and AUC, while keeping a good balance between the sensitivity and specificity. Results demonstrate great potentials of the proposed CNN-based approach for the prediction of MCI-to-AD conversion with solely MRI data. Age correction and assisted structural brain image features can boost the prediction performance of CNN.

Learning Pairwise-Similarity Guided Sparse Functional Connectivity Network for MCI Classification

Learning functional connectivity (FC) network from resting-state function magnetic resonance imaging (RS-fMRI) data via sparse representation (SR) or weighted SR (WSR) has been proved to be promising for the diagnosis of Alzheimer's disease and its prodromal stage, mild cognitive impairment (MCI). However, traditional SR/WSR based approaches learn the representation of each brain region independently, without fully taking into account the possible relationship between brain regions. To remedy this limitation, we propose a novel FC modeling approach by considering two types of possible relationship between different brain regions which are incorporated into SR/WSR approaches in the form of regularization. In this way, the representations of all brain regions can be jointly learned. Furthermore, an efficient alternating optimization algorithm is also developed to solve the resulting model. Experimental results show that our proposed method not only outperforms SR and WSR in the diagnosis of MCI subjects, but also leads to the brain FC network with better modularity structure.

Pixel-Label-Based Segmentation of Cross-Sectional Brain MRI Using Simplified SegNet Architecture-Based CNN

Using deep neural networks for segmenting an MRI image of heterogeneously distributed pixels into a specific class assigning a label to each pixel is the concept of the proposed approach. This approach facilitates the application of the segmentation process on a preprocessed MRI image, with a trained network to be utilized for other test images. As labels are considered expensive assets in supervised training, fewer training images and training labels are used to obtain optimal accuracy. To validate the performance of the proposed approach, an experiment is conducted on other test images (available in the same database) that are not part of the training; the obtained result is of good visual quality in terms of segmentation and quite similar to the ground truth image. The average computed Dice similarity index for the test images is approximately 0.8, whereas the Jaccard similarity measure is approximately 0.6, which is better compared to other methods. This implies that the proposed method can be used to obtain reference images almost similar to the segmented ground truth images.

SEMI-SUPERVISED LEARNING FOR PELVIC MR IMAGE SEGMENTATION BASED ON MULTI-TASK RESIDUAL FULLY CONVOLUTIONAL NETWORKS

Accurate segmentation of pelvic organs from magnetic resonance (MR) images plays an important role in image-guided radiotherapy. However, it is a challenging task due to inconsistent organ appearances and large shape variations. Fully convolutional network (FCN) has recently achieved state-of-the-art performance in medical image segmentation, but it requires a large amount of labeled data for training, which is usually difficult to obtain in real situation. To address these challenges, we propose a deep learning based semi-supervised learning framework. Specifically, we first train an initial multi-task residual fully convolutional network (FCN) based on a limited number of labeled MRI data. Based on the initially trained FCN, those unlabeled new data can be automatically segmented and some reasonable segmentations (after manual/automatic checking) can be included into the training data to fine-tune the network. This step can be repeated to progressively improve the training of our network, until no reasonable segmentations of new data can be included. Experimental results demonstrate the effectiveness of our proposed progressive semi-supervised learning fashion as well as its advantage in terms of accuracy.

Computational neuroanatomy of baby brains: A review

The first postnatal years are an exceptionally dynamic and critical period of structural, functional and connectivity development of the human brain. The increasing availability of non-invasive infant brain MR images provides unprecedented opportunities for accurate and reliable charting of dynamic early brain developmental trajectories in understanding normative and aberrant growth. However, infant brain MR images typically exhibit reduced tissue contrast (especially around 6 months of age), large within-tissue intensity variations, and regionally-heterogeneous, dynamic changes, in comparison with adult brain MR images. Consequently, the existing computational tools developed typically for adult brains are not suitable for infant brain MR image processing. To address these challenges, many infant-tailored computational methods have been proposed for computational neuroanatomy of infant brains. In this review paper, we provide a comprehensive review of the state-of-the-art computational methods for infant brain MRI processing and analysis, which have advanced our understanding of early postnatal brain development. We also summarize publically available infant-dedicated resources, including MRI datasets, computational tools, grand challenges, and brain atlases. Finally, we discuss the limitations in current research and suggest potential future research directions.

Segmentation of Craniomaxillofacial Bony Structures from MRI with a 3D Deep-Learning Based Cascade Framework

Computed tomography (CT) is commonly used as a diagnostic and treatment planning imaging modality in craniomaxillofacial (CMF) surgery to correct patient's bony defects. A major disadvantage of CT is that it emits harmful ionizing radiation to patients during the exam. Magnetic resonance imaging (MRI) is considered to be much safer and noninvasive, and often used to study CMF soft tissues (e.g., temporomandibular joint and brain). However, it is extremely difficult to accurately segment CMF bony structures from MRI since both bone and air appear to be black in MRI, along with low signal-to-noise ratio and partial volume effect. To this end, we proposed a 3D deep-learning based cascade framework to solve these issues. Specifically, a 3D fully convolutional network (FCN) architecture is first adopted to coarsely segment the bony structures. As the coarsely segmented bony structures by FCN tend to be thicker, convolutional neural network (CNN) is further utilized for fine-grained segmentation. To enhance the discriminative ability of the CNN, we particularly concatenate the predicted probability maps from FCN and the original MRI, and feed them together into the CNN to provide more context information for segmentation. Experimental results demonstrate a good performance and also the clinical feasibility of our proposed 3D deep-learning based cascade framework.

Medical Image Synthesis with Context-Aware Generative Adversarial Networks

Computed tomography (CT) is critical for various clinical applications, e.g., radiation treatment planning and also PET attenuation correction in MRI/PET scanner. However, CT exposes radiation during acquisition, which may cause side effects to patients. Compared to CT, magnetic resonance imaging (MRI) is much safer and does not involve radiations. Therefore, recently researchers are greatly motivated to estimate CT image from its corresponding MR image of the same subject for the case of radiation planning. In this paper, we propose a data-driven approach to address this challenging problem. Specifically, we train a fully convolutional network (FCN) to generate CT given the MR image. To better model the nonlinear mapping from MRI to CT and produce more realistic images, we propose to use the adversarial training strategy to train the FCN. Moreover, we propose an image-gradient-difference based loss function to alleviate the blurriness of the generated CT. We further apply Auto-Context Model (ACM) to implement a context-aware generative adversarial network. Experimental results show that our method is accurate and robust for predicting CT images from MR images, and also outperforms three state-of-the-art methods under comparison.

Joint Segmentation of Multiple Thoracic Organs in CT Images with Two Collaborative Deep Architectures

Computed Tomography (CT) is the standard imaging technique for radiotherapy planning. The delineation of Organs at Risk (OAR) in thoracic CT images is a necessary step before radiotherapy, for preventing irradiation of healthy organs. However, due to low contrast, multi-organ segmentation is a challenge. In this paper, we focus on developing a novel framework for automatic delineation of OARs. Different from previous works in OAR segmentation where each organ is segmented separately, we propose two collaborative deep architectures to jointly segment all organs, including esophagus, heart, aorta and trachea. Since most of the organ borders are ill-defined, we believe spatial relationships must be taken into account to overcome the lack of contrast. The aim of combining two networks is to learn anatomical constraints with the first network, which will be used in the second network, when each OAR is segmented in turn. Specifically, we use the first deep architecture, a deep SharpMask architecture, for providing an effective combination of low-level representations with deep high-level features, and then take into account the spatial relationships between organs by the use of Conditional Random Fields (CRF). Next, the second deep architecture is employed to refine the segmentation of each organ by using the maps obtained on the first deep architecture to learn anatomical constraints for guiding and refining the segmentations. Experimental results show superior performance on 30 CT scans, comparing with other state-of-the-art methods.

Brain Image Labeling Using Multi-atlas Guided 3D Fully Convolutional Networks

Automatic labeling of anatomical structures in brain images plays an important role in neuroimaging analysis. Among all methods, multi-atlas based segmentation methods are widely used, due to their robustness in propagating prior label information. However, non-linear registration is always needed, which is time-consuming. Alternatively, the patch-based methods have been proposed to relax the requirement of image registration, but the labeling is often determined independently by the target image information, without getting direct assistance from the atlases. To address these limitations, in this paper, we propose a multi-atlas guided 3D fully convolutional networks (FCN) for brain image labeling. Specifically, multi-atlas based guidance is incorporated during the network learning. Based on this, the discriminative of the FCN is boosted, which eventually contribute to accurate prediction. Experiments show that the use of multi-atlas guidance improves the brain labeling performance.

Deep Learning for Brain MRI Segmentation: State of the Art and Future Directions

Quantitative analysis of brain MRI is routine for many neurological diseases and conditions and relies on accurate segmentation of structures of interest. Deep learning-based segmentation approaches for brain MRI are gaining interest due to their self-learning and generalization ability over large amounts of data. As the deep learning architectures are becoming more mature, they gradually outperform previous state-of-the-art classical machine learning algorithms. This review aims to provide an overview of current deep learning-based segmentation approaches for quantitative brain MRI. First we review the current deep learning architectures used for segmentation of anatomical brain structures and brain lesions. Next, the performance, speed, and properties of deep learning approaches are summarized and discussed. Finally, we provide a critical assessment of the current state and identify likely future developments and trends.

Location Sensitive Deep Convolutional Neural Networks for Segmentation of White Matter Hyperintensities

The anatomical location of imaging features is of crucial importance for accurate diagnosis in many medical tasks. Convolutional neural networks (CNN) have had huge successes in computer vision, but they lack the natural ability to incorporate the anatomical location in their decision making process, hindering success in some medical image analysis tasks. In this paper, to integrate the anatomical location information into the network, we propose several deep CNN architectures that consider multi-scale patches or take explicit location features while training. We apply and compare the proposed architectures for segmentation of white matter hyperintensities in brain MR images on a large dataset. As a result, we observe that the CNNs that incorporate location information substantially outperform a conventional segmentation method with handcrafted features as well as CNNs that do not integrate location information. On a test set of 50 scans, the best configuration of our networks obtained a Dice score of 0.792, compared to 0.805 for an independent human observer. Performance levels of the machine and the independent human observer were not statistically significantly different (p-value = 0.06).

Deep Learning in Medical Image Analysis

This review covers computer-assisted analysis of images in the field of medical imaging. Recent advances in machine learning, especially with regard to deep learning, are helping to identify, classify, and quantify patterns in medical images. At the core of these advances is the ability to exploit hierarchical feature representations learned solely from data, instead of features designed by hand according to domain-specific knowledge. Deep learning is rapidly becoming the state of the art, leading to enhanced performance in various medical applications. We introduce the fundamentals of deep learning methods and review their successes in image registration, detection of anatomical and cellular structures, tissue segmentation, computer-aided disease diagnosis and prognosis, and so on. We conclude by discussing research issues and suggesting future directions for further improvement.

Integrating Online and Offline Three-Dimensional Deep Learning for Automated Polyp Detection in Colonoscopy Videos

Automated polyp detection in colonoscopy videos has been demonstrated to be a promising way for colorectal cancer prevention and diagnosis. Traditional manual screening is time consuming, operator dependent, and error prone; hence, automated detection approach is highly demanded in clinical practice. However, automated polyp detection is very challenging due to high intraclass variations in polyp size, color, shape, and texture, and low interclass variations between polyps and hard mimics. In this paper, we propose a novel offline and online three-dimensional (3-D) deep learning integration framework by leveraging the 3-D fully convolutional network (3D-FCN) to tackle this challenging problem. Compared with the previous methods employing hand-crafted features or 2-D convolutional neural network, the 3D-FCN is capable of learning more representative spatio-temporal features from colonoscopy videos, and hence has more powerful discrimination capability. More importantly, we propose a novel online learning scheme to deal with the problem of limited training data by harnessing the specific information of an input video in the learning process. We integrate offline and online learning to effectively reduce the number of false positives generated by the offline network and further improve the detection performance. Extensive experiments on the dataset of MICCAI 2015 Challenge on Polyp Detection demonstrated the better performance of our method when compared with other competitors.

Estimating CT Image from MRI Data Using 3D Fully Convolutional Networks

Computed tomography (CT) is critical for various clinical applications, e.g., radiotherapy treatment planning and also PET attenuation correction. However, CT exposes radiation during CT imaging, which may cause side effects to patients. Compared to CT, magnetic resonance imaging (MRI) is much safer and does not involve any radiation. Therefore, recently researchers are greatly motivated to estimate CT image from its corresponding MR image of the same subject for the case of radiotherapy planning. In this paper, we propose a 3D deep learning based method to address this challenging problem. Specifically, a 3D fully convolutional neural network (FCN) is adopted to learn an end-to-end nonlinear mapping from MR image to CT image. Compared to the conventional convolutional neural network (CNN), FCN generates structured output and can better preserve the neighborhood information in the predicted CT image. We have validated our method in a real pelvic CT/MRI dataset. Experimental results show that our method is accurate and robust for predicting CT image from MRI image, and also outperforms three state-of-the-art methods under comparison. In addition, the parameters, such as network depth and activation function, are extensively studied to give an insight for deep learning based regression tasks in our application.