Boundary-Weighted Domain Adaptive Neural Network for Prostate MR Image Segmentation

Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation

Automatic segmentation methods are an important advancement in medical image analysis. Machine learning techniques, and deep neural networks in particular, are the state-of-the-art for most medical image segmentation tasks. Issues with class imbalance pose a significant challenge in medical datasets, with lesions often occupying a considerably smaller volume relative to the background. Loss functions used in the training of deep learning algorithms differ in their robustness to class imbalance, with direct consequences for model convergence. The most commonly used loss functions for segmentation are based on either the cross entropy loss, Dice loss or a combination of the two. We propose the Unified Focal loss, a new hierarchical framework that generalises Dice and cross entropy-based losses for handling class imbalance. We evaluate our proposed loss function on five publicly available, class imbalanced medical imaging datasets: CVC-ClinicDB, Digital Retinal Images for Vessel Extraction (DRIVE), Breast Ultrasound 2017 (BUS2017), Brain Tumour Segmentation 2020 (BraTS20) and Kidney Tumour Segmentation 2019 (KiTS19). We compare our loss function performance against six Dice or cross entropy-based loss functions, across 2D binary, 3D binary and 3D multiclass segmentation tasks, demonstrating that our proposed loss function is robust to class imbalance and consistently outperforms the other loss functions. Source code is available at: https://github.com/mlyg/unified-focal-loss.

•

Loss function choice is crucial for class-imbalanced medical imaging datasets.

•

Understanding the relationship between loss functions is key to inform choice.

•

Unified Focal loss generalises Dice and cross-entropy based loss functions.

•

Unified Focal loss outperforms various Dice and cross-entropy based loss functions.

Dynamic boundary-insensitive loss for magnetic resonance medical image segmentation

PURPOSE

A deep learning method has achieved great success in MR medical image segmentation. One challenge in applying deep learning segmentation models to clinical practice is their poor generalization mainly due to limited labeled training samples, inter-site heterogeneity of different datasets, and ambiguous boundary definition, etc. The objective of this work is to develop a dynamic boundary-insensitive (DBI) loss to address this poor generalization caused by an uncertain boundary.

METHODS

The DBI loss is designed to assign higher penalties to misclassified voxels farther from the boundaries in each training iteration to reduce the sensitivity of the segmentation model to the uncertain boundary. The weighting factor of the DBI loss can be adjusted adaptively without any manual setting and adjustment. Extensive experiments were conducted to verify the performance of our DBI loss and its variant, DiceDBI, on four heterogeneous prostate MRI datasets for prostate zonal segmentation and whole prostate segmentation.

RESULTS

Experimental results show that our DBI loss, when combined with Dice loss, outperforms all competing loss functions in dice similarity coefficient (DSC) and improves the segmentation performance across all datasets consistently, especially on unseen datasets and when segmenting small or narrow targets.

CONCLUSIONS

The proposed DiceDBI loss will be valuable for enhancement of the generalization performance of the segmentation model.

Deep Hybrid Convolutional Neural Network for Segmentation of Melanoma Skin Lesion

Melanoma is a type of skin cancer that often leads to poor prognostic responses and survival rates. Melanoma usually develops in the limbs, including in fingers, palms, and the margins of the nails. When melanoma is detected early, surgical treatment may achieve a higher cure rate. The early diagnosis of melanoma depends on the manual segmentation of suspected lesions. However, manual segmentation can lead to problems, including misclassification and low efficiency. Therefore, it is essential to devise a method for automatic image segmentation that overcomes the aforementioned issues. In this study, an improved algorithm is proposed, termed EfficientUNet++, which is developed from the U-Net model. In EfficientUNet++, the pretrained EfficientNet model is added to the UNet++ model to accelerate segmentation process, leading to more reliable and precise results in skin cancer image segmentation. Two skin lesion datasets were used to compare the performance of the proposed EfficientUNet++ algorithm with other common models. In the PH2 dataset, EfficientUNet++ achieved a better Dice coefficient (93% vs. 76%-91%), Intersection over Union (IoU, 96% vs. 74%-95%), and loss value (30% vs. 44%-32%) compared with other models. In the International Skin Imaging Collaboration dataset, EfficientUNet++ obtained a similar Dice coefficient (96% vs. 94%-96%) but a better IoU (94% vs. 89%-93%) and loss value (11% vs. 13%-11%) than other models. In conclusion, the EfficientUNet++ model efficiently detects skin lesions by improving composite coefficients and structurally expanding the size of the convolution network. Moreover, the use of residual units deepens the network to further improve performance.

Automatic Segmentation of Pelvic Cancers Using Deep Learning: State-of-the-Art Approaches and Challenges

The recent rise of deep learning (DL) and its promising capabilities in capturing non-explicit detail from large datasets have attracted substantial research attention in the field of medical image processing. DL provides grounds for technological development of computer-aided diagnosis and segmentation in radiology and radiation oncology. Amongst the anatomical locations where recent auto-segmentation algorithms have been employed, the pelvis remains one of the most challenging due to large intra- and inter-patient soft-tissue variabilities. This review provides a comprehensive, non-systematic and clinically-oriented overview of 74 DL-based segmentation studies, published between January 2016 and December 2020, for bladder, prostate, cervical and rectal cancers on computed tomography (CT) and magnetic resonance imaging (MRI), highlighting the key findings, challenges and limitations.

Outbreak COVID-19 in Medical Image Processing Using Deep Learning: A State-of-the-Art Review

From the month of December-19, the outbreak of Coronavirus (COVID-19) triggered several deaths and overstated every aspect of individual health. COVID-19 has been designated as a pandemic by World Health Organization. The circumstances placed serious trouble on every country worldwide, particularly with health arrangements and time-consuming responses. The increase in the positive cases of COVID-19 globally spread every day. The quantity of accessible diagnosing kits is restricted because of complications in detecting the existence of the illness. Fast and correct diagnosis of COVID-19 is a timely requirement for the prevention and controlling of the pandemic through suitable isolation and medicinal treatment. The significance of the present work is to discuss the outline of the deep learning techniques with medical imaging such as outburst prediction, virus transmitted indications, detection and treatment aspects, vaccine availability with remedy research. Abundant image resources of medical imaging as X-rays, Computed Tomography Scans, Magnetic Resonance imaging, formulate deep learning high-quality methods to fight against the pandemic COVID-19. The review presents a comprehensive idea of deep learning and its related applications in healthcare received over the past decade. At the last, some issues and confrontations to control the health crisis and outbreaks have been introduced. The progress in technology has contributed to developing individual's lives. The problems faced by the radiologists during medical imaging techniques and deep learning approaches for diagnosing the COVID-19 infections have been also discussed.

Domain adaptation based self-correction model for COVID-19 infection segmentation in CT images

The capability of generalization to unseen domains is crucial for deep learning models when considering real-world scenarios. However, current available medical image datasets, such as those for COVID-19 CT images, have large variations of infections and domain shift problems. To address this issue, we propose a prior knowledge driven domain adaptation and a dual-domain enhanced self-correction learning scheme. Based on the novel learning scheme, a domain adaptation based self-correction model (DASC-Net) is proposed for COVID-19 infection segmentation on CT images. DASC-Net consists of a novel attention and feature domain enhanced domain adaptation model (AFD-DA) to solve the domain shifts and a self-correction learning process to refine segmentation results. The innovations in AFD-DA include an image-level activation feature extractor with attention to lung abnormalities and a multi-level discrimination module for hierarchical feature domain alignment. The proposed self-correction learning process adaptively aggregates the learned model and corresponding pseudo labels for the propagation of aligned source and target domain information to alleviate the overfitting to noises caused by pseudo labels. Extensive experiments over three publicly available COVID-19 CT datasets demonstrate that DASC-Net consistently outperforms state-of-the-art segmentation, domain shift, and coronavirus infection segmentation methods. Ablation analysis further shows the effectiveness of the major components in our model. The DASC-Net enriches the theory of domain adaptation and self-correction learning in medical imaging and can be generalized to multi-site COVID-19 infection segmentation on CT images for clinical deployment.

Versatile Convolutional Networks Applied to Computed Tomography and Magnetic Resonance Image Segmentation

Medical image segmentation has seen positive developments in recent years but remains challenging with many practical obstacles to overcome. The applications of this task are wide-ranging in many fields of medicine, and used in several imaging modalities which usually require tailored solutions. Deep learning models have gained much attention and have been lately recognized as the most successful for automated segmentation. In this work we show the versatility of this technique by means of a single deep learning architecture capable of successfully performing segmentation on two very different types of imaging: computed tomography and magnetic resonance. The developed model is fully convolutional with an encoder-decoder structure and high-resolution pathways which can process whole three-dimensional volumes at once, and learn directly from the data to find which voxels belong to the regions of interest and localize those against the background. The model was applied to two publicly available datasets achieving equivalent results for both imaging modalities, as well as performing segmentation of different organs in different anatomic regions with comparable success.

Contrastive rendering with semi-supervised learning for ovary and follicle segmentation from 3D ultrasound

Segmentation of ovary and follicles from 3D ultrasound (US) is the crucial technique of measurement tools for female infertility diagnosis. Since manual segmentation is time-consuming and operator-dependent, an accurate and fast segmentation method is highly demanded. However, it is challenging for current deep-learning based methods to segment ovary and follicles precisely due to ambiguous boundaries and insufficient annotations. In this paper, we propose a contrastive rendering (C-Rend) framework to segment ovary and follicles with detail-refined boundaries. Furthermore, we incorporate the proposed C-Rend with a semi-supervised learning (SSL) framework, leveraging unlabeled data for better performance. Highlights of this paper include: (1) A rendering task is performed to estimate boundary accurately via enriched feature representation learning. (2) Point-wise contrastive learning is proposed to enhance the similarity of intra-class points and contrastively decrease the similarity of inter-class points. (3) The C-Rend plays a complementary role for the SSL framework in uncertainty-aware learning, which could provide reliable supervision information and achieve superior segmentation performance. Through extensive validation on large in-house datasets with partial annotations, our method outperforms state-of-the-art methods in various evaluation metrics for both the ovary and follicles.

Automated cardiac segmentation of cross-modal medical images using unsupervised multi-domain adaptation and spatial neural attention structure

Accurate cardiac segmentation of multimodal images, e.g., magnetic resonance (MR), computed tomography (CT) images, plays a pivot role in auxiliary diagnoses, treatments and postoperative assessments of cardiovascular diseases. However, training a well-behaved segmentation model for the cross-modal cardiac image analysis is challenging, due to their diverse appearances/distributions from different devices and acquisition conditions. For instance, a well-trained segmentation model based on the source domain of MR images is often failed in the segmentation of CT images. In this work, a cross-modal images-oriented cardiac segmentation scheme is proposed using a symmetric full convolutional neural network (SFCNN) with the unsupervised multi-domain adaptation (UMDA) and a spatial neural attention (SNA) structure, termed UMDA-SNA-SFCNN, having the merits of without the requirement of any annotation on the test domain. Specifically, UMDA-SNA-SFCNN incorporates SNA to the classic adversarial domain adaptation network to highlight the relevant regions, while restraining the irrelevant areas in the cross-modal images, so as to suppress the negative transfer in the process of unsupervised domain adaptation. In addition, the multi-layer feature discriminators and a predictive segmentation-mask discriminator are established to connect the multi-layer features and segmentation mask of the backbone network, SFCNN, to realize the fine-grained alignment of unsupervised cross-modal feature domains. Extensive confirmative and comparative experiments on the benchmark Multi-Modality Whole Heart Challenge dataset show that the proposed model is superior to the state-of-the-art cross-modal segmentation methods.

Domain adaptation for segmentation of critical structures for prostate cancer therapy

Preoperative assessment of the proximity of critical structures to the tumors is crucial in avoiding unnecessary damage during prostate cancer treatment. A patient-specific 3D anatomical model of those structures, namely the neurovascular bundles (NVB) and the external urethral sphincters (EUS), can enable physicians to perform such assessments intuitively. As a crucial step to generate a patient-specific anatomical model from preoperative MRI in a clinical routine, we propose a multi-class automatic segmentation based on an anisotropic convolutional network. Our specific challenge is to train the network model on a unique source dataset only available at a single clinical site and deploy it to another target site without sharing the original images or labels. As network models trained on data from a single source suffer from quality loss due to the domain shift, we propose a semi-supervised domain adaptation (DA) method to refine the model's performance in the target domain. Our DA method combines transfer learning and uncertainty guided self-learning based on deep ensembles. Experiments on the segmentation of the prostate, NVB, and EUS, show significant performance gain with the combination of those techniques compared to pure TL and the combination of TL with simple self-learning ([Formula: see text] for all structures using a Wilcoxon's signed-rank test). Results on a different task and data (Pancreas CT segmentation) demonstrate our method's generic application capabilities. Our method has the advantage that it does not require any further data from the source domain, unlike the majority of recent domain adaptation strategies. This makes our method suitable for clinical applications, where the sharing of patient data is restricted.

Dual temporal gated multi-graph convolution network for taxi demand prediction

Taxi demand prediction is essential to build efficient traffic transportation systems for smart city. It helps to properly allocate vehicles, ease the traffic pressure and improve passengers’ experience. Traditional taxi demand prediction methods mostly rely on time-series forecasting techniques, which cannot model the nonlinearity embedded in data. Recent studies start to combine the Euclidean spatial features through grid-based methods. By considering the spatial correlations among different regions, we can capture how the temporal events have impacts on those with adjacent links or intersections and improve prediction precision. Some graph-based models are proposed to encode the non-Euclidean correlations as well. However, the temporal periodicity of data is often overlooked, and the study units are usually constructed as oversimplified grids. In this paper, we define places with specific semantic and humanistic experiences as study units, using a fuzzy set method based on adaptive kernel density estimation. Then, we introduce dual temporal gated multi-graph convolution network to predict the future taxi demand. Specifically, multi-graph convolution is used to model spatial correlations with graphs, including the neighborhood, functional similarities and landscape similarities based on street view images. As for the temporal dependencies modeling, we design the dual temporal gated branches to capture information hidden in both previous and periodic observations. Experiments on two real-world datasets show the effectiveness of our model over the baselines.

Dual temporal gated multi-graph convolution network for taxi demand prediction

Taxi demand prediction is essential to build efficient traffic transportation systems for smart city. It helps to properly allocate vehicles, ease the traffic pressure and improve passengers’ experience. Traditional taxi demand prediction methods mostly rely on time-series forecasting techniques, which cannot model the nonlinearity embedded in data. Recent studies start to combine the Euclidean spatial features through grid-based methods. By considering the spatial correlations among different regions, we can capture how the temporal events have impacts on those with adjacent links or intersections and improve prediction precision. Some graph-based models are proposed to encode the non-Euclidean correlations as well. However, the temporal periodicity of data is often overlooked, and the study units are usually constructed as oversimplified grids. In this paper, we define places with specific semantic and humanistic experiences as study units, using a fuzzy set method based on adaptive kernel density estimation. Then, we introduce dual temporal gated multi-graph convolution network to predict the future taxi demand. Specifically, multi-graph convolution is used to model spatial correlations with graphs, including the neighborhood, functional similarities and landscape similarities based on street view images. As for the temporal dependencies modeling, we design the dual temporal gated branches to capture information hidden in both previous and periodic observations. Experiments on two real-world datasets show the effectiveness of our model over the baselines.

Multi-Task Learning for Registering Images With Large Deformation

Accurate registration of prostate magnetic resonance imaging (MRI) images of the same subject acquired at different time points helps diagnose cancer and monitor the tumor progress. However, it is very challenging especially when one image was acquired with the use of endorectal coil (ERC) but the other was not, which causes significant deformation. Classical iterative image registration methods are also computationally intensive. Deep learning based registration frameworks have recently been developed and demonstrated promising performance. However, the lack of proper constraints often results in unrealistic registration. In this paper, we propose a multi-task learning based registration network with anatomical constraint to address these issues. The proposed approach uses a cycle constraint loss to achieve forward/backward registration and an inverse constraint loss to encourage diffeomorphic registration. In addition, an adaptive anatomical constraint aiming for regularizing the registration network with the use of anatomical labels is introduced through weak supervision. Our experiments on registering prostate MR images of the same subject obtained at different time points with and without ERC show that the proposed method achieves very promising performance under different measures in dealing with the large deformation. Compared with other existing methods, our approach works more efficiently with average running time less than a second and is able to obtain more visually realistic results.

Deep learning and medical image processing for coronavirus (COVID-19) pandemic: A survey

Since December 2019, the coronavirus disease (COVID-19) outbreak has caused many death cases and affected all sectors of human life. With gradual progression of time, COVID-19 was declared by the world health organization (WHO) as an outbreak, which has imposed a heavy burden on almost all countries, especially ones with weaker health systems and ones with slow responses. In the field of healthcare, deep learning has been implemented in many applications, e.g., diabetic retinopathy detection, lung nodule classification, fetal localization, and thyroid diagnosis. Numerous sources of medical images (e.g., X-ray, CT, and MRI) make deep learning a great technique to combat the COVID-19 outbreak. Motivated by this fact, a large number of research works have been proposed and developed for the initial months of 2020. In this paper, we first focus on summarizing the state-of-the-art research works related to deep learning applications for COVID-19 medical image processing. Then, we provide an overview of deep learning and its applications to healthcare found in the last decade. Next, three use cases in China, Korea, and Canada are also presented to show deep learning applications for COVID-19 medical image processing. Finally, we discuss several challenges and issues related to deep learning implementations for COVID-19 medical image processing, which are expected to drive further studies in controlling the outbreak and controlling the crisis, which results in smart healthy cities.

A Tour of Unsupervised Deep Learning for Medical Image Analysis

BACKGROUND

Interpretation of medical images for the diagnosis and treatment of complex diseases from high-dimensional and heterogeneous data remains a key challenge in transforming healthcare. In the last few years, both supervised and unsupervised deep learning achieved promising results in the area of medical image analysis. Several reviews on supervised deep learning are published, but hardly any rigorous review on unsupervised deep learning for medical image analysis is available.

OBJECTIVES

The objective of this review is to systematically present various unsupervised deep learning models, tools, and benchmark datasets applied to medical image analysis. Some of the discussed models are autoencoders and its other variants, Restricted Boltzmann machines (RBM), Deep belief networks (DBN), Deep Boltzmann machine (DBM), and Generative adversarial network (GAN). Further, future research opportunities and challenges of unsupervised deep learning techniques for medical image analysis are also discussed.

CONCLUSION

Currently, interpretation of medical images for diagnostic purposes is usually performed by human experts that may be replaced by computer-aided diagnosis due to advancement in machine learning techniques, including deep learning, and the availability of cheap computing infrastructure through cloud computing. Both supervised and unsupervised machine learning approaches are widely applied in medical image analysis, each of them having certain pros and cons. Since human supervisions are not always available or inadequate or biased, therefore, unsupervised learning algorithms give a big hope with lots of advantages for biomedical image analysis.

Multi-task edge-recalibrated network for male pelvic multi-organ segmentation on CT images

Automated male pelvic multi-organ segmentation on CT images is highly desired for applications, including radiotherapy planning. To further improve the performance and efficiency of existing automated segmentation methods, in this study, we propose a multi-task edge-recalibrated network (MTER-Net), which aims to overcome the challenges, including blurry boundaries, large inter-patient appearance variations, and low soft-tissue contrast. The proposed MTER-Net is equipped with the following novel components. (a) To exploit the saliency and stability of femoral heads, we employed a light-weight localization module to locate the target region and efficiently remove the complex background. (b) We add an edge stream to the regular segmentation stream to focus on processing the edge-related information, distinguish the organs with blurry boundaries, and then boost the overall segmentation performance. Between the regular segmentation stream and edge stream, we introduce an edge recalibration module at each resolution level to connect the intermediate layers and deliver the higher-level activations from the regular stream to the edge stream to denoise the irrelevant activations. (c) Finally, using a 3D Atrous Spatial Pyramid Pooling (ASPP) feature fusion module, we fuse the features at different scales in the regular stream and the predictions from the edge stream to form the final segmentation result. The proposed segmentation network was evaluated on 200 prostate cancer patient CT images with manually delineated contours of bladder, rectum, seminal vesicle, and prostate. The segmentation performance of the proposed method was quantitatively evaluated using three metrics including Dice similarity coefficient (DSC), average surface distance (ASD), and 95% surface distance (95SD). The proposed MTER-Net achieves average DSC of 86.35%, ASD of 1.09 mm, and 95SD of 3.53 mm on the four organs, which outperforms the state-of-the-art segmentation networks by a large margin. Specifically, the quantitative DSC evaluation results of the four organs are 96.49% (bladder), 86.39% (rectum), 76.38% (seminal vesicle), and 86.14% (prostate), respectively. In conclusion, we demonstrate that the proposed MTER-Net efficiently attains superior performance to state-of-the-art pelvic organ segmentation methods.

3D PBV-Net: An automated prostate MRI data segmentation method

Prostate cancer is one of the most common deadly diseases in men worldwide, which is seriously affecting people's life and health. Reliable and automated segmentation of the prostate gland in MRI data is exceptionally critical for diagnosis and treatment planning of prostate cancer. Although many automated segmentation methods have emerged, including deep learning based approaches, segmentation performance is still poor due to the large variability of image appearance, anisotropic spatial resolution, and imaging interference. This study proposes an automated prostate MRI data segmentation approach using bicubic interpolation with improved 3D V-Net (dubbed 3D PBV-Net). Considering the low-frequency components in the prostate gland, the bicubic interpolation is applied to preprocess the MRI data. On this basis, a 3D PBV-Net is developed to perform prostate MRI data segmentation. To illustrate the effectiveness of our approach, we evaluate the proposed 3D PBV-Net on two clinical prostate MRI data datasets, i.e., PROMISE 12 and TPHOH, with the manual delineations available as the ground truth. Our approach generates promising segmentation results, which have achieved 97.65% and 98.29% of average accuracy, 0.9613 and 0.9765 of Dice metric, 3.120 mm and 0.9382 mm of Hausdorff distance, and average boundary distance of 1.708, 0.7950 on PROMISE 12 and TPHOH datasets, respectively. Our method has effectively improved the accuracy of automated segmentation of the prostate MRI data and is promising to meet the accuracy requirements for telehealth applications.

PSIGAN: Joint Probabilistic Segmentation and Image Distribution Matching for Unpaired Cross-Modality Adaptation-Based MRI Segmentation

We developed a new joint probabilistic segmentation and image distribution matching generative adversarial network (PSIGAN) for unsupervised domain adaptation (UDA) and multi-organ segmentation from magnetic resonance (MRI) images. Our UDA approach models the co-dependency between images and their segmentation as a joint probability distribution using a new structure discriminator. The structure discriminator computes structure of interest focused adversarial loss by combining the generated pseudo MRI with probabilistic segmentations produced by a simultaneously trained segmentation sub-network. The segmentation sub-network is trained using the pseudo MRI produced by the generator sub-network. This leads to a cyclical optimization of both the generator and segmentation sub-networks that are jointly trained as part of an end-to-end network. Extensive experiments and comparisons against multiple state-of-the-art methods were done on four different MRI sequences totalling 257 scans for generating multi-organ and tumor segmentation. The experiments included, (a) 20 T1-weighted (T1w) in-phase mdixon and (b) 20 T2-weighted (T2w) abdominal MRI for segmenting liver, spleen, left and right kidneys, (c) 162 T2-weighted fat suppressed head and neck MRI (T2wFS) for parotid gland segmentation, and (d) 75 T2w MRI for lung tumor segmentation. Our method achieved an overall average DSC of 0.87 on T1w and 0.90 on T2w for the abdominal organs, 0.82 on T2wFS for the parotid glands, and 0.77 on T2w MRI for lung tumors.

Application of Artificial Intelligence Technology in Oncology: Towards the Establishment of Precision Medicine

In recent years, advances in artificial intelligence (AI) technology have led to the rapid clinical implementation of devices with AI technology in the medical field. More than 60 AI-equipped medical devices have already been approved by the Food and Drug Administration (FDA) in the United States, and the active introduction of AI technology is considered to be an inevitable trend in the future of medicine. In the field of oncology, clinical applications of medical devices using AI technology are already underway, mainly in radiology, and AI technology is expected to be positioned as an important core technology. In particular, "precision medicine," a medical treatment that selects the most appropriate treatment for each patient based on a vast amount of medical data such as genome information, has become a worldwide trend; AI technology is expected to be utilized in the process of extracting truly useful information from a large amount of medical data and applying it to diagnosis and treatment. In this review, we would like to introduce the history of AI technology and the current state of medical AI, especially in the oncology field, as well as discuss the possibilities and challenges of AI technology in the medical field.

Multi-Organ Segmentation Over Partially Labeled Datasets With Multi-Scale Feature Abstraction

Shortage of fully annotated datasets has been a limiting factor in developing deep learning based image segmentation algorithms and the problem becomes more pronounced in multi-organ segmentation. In this paper, we propose a unified training strategy that enables a novel multi-scale deep neural network to be trained on multiple partially labeled datasets for multi-organ segmentation. In addition, a new network architecture for multi-scale feature abstraction is proposed to integrate pyramid input and feature analysis into a U-shape pyramid structure. To bridge the semantic gap caused by directly merging features from different scales, an equal convolutional depth mechanism is introduced. Furthermore, we employ a deep supervision mechanism to refine the outputs in different scales. To fully leverage the segmentation features from all the scales, we design an adaptive weighting layer to fuse the outputs in an automatic fashion. All these mechanisms together are integrated into a Pyramid Input Pyramid Output Feature Abstraction Network (PIPO-FAN). Our proposed method was evaluated on four publicly available datasets, including BTCV, LiTS, KiTS and Spleen, where very promising performance has been achieved. The source code of this work is publicly shared at https://github.com/DIAL-RPI/PIPO-FAN to facilitate others to reproduce the work and build their own models using the introduced mechanisms.

Adversarial Confidence Learning for Medical Image Segmentation and Synthesis

Generative adversarial networks (GAN) are widely used in medical image analysis tasks, such as medical image segmentation and synthesis. In these works, adversarial learning is directly applied to the original supervised segmentation (synthesis) networks. The usage of adversarial learning is effective in improving visual perception performance since adversarial learning works as realistic regularization for supervised generators. However, the quantitative performance often cannot improve as much as the qualitative performance, and it can even become worse in some cases. In this paper, we explore how we can take better advantage of adversarial learning in supervised segmentation (synthesis) models and propose an adversarial confidence learning framework to better model these problems. We analyze the roles of discriminator in the classic GANs and compare them with those in supervised adversarial systems. Based on this analysis, we propose adversarial confidence learning, i.e., besides the adversarial learning for emphasizing visual perception, we use the confidence information provided by the adversarial network to enhance the design of supervised segmentation (synthesis) network. In particular, we propose using a fully convolutional adversarial network for confidence learning to provide voxel-wise and region-wise confidence information for the segmentation (synthesis) network. With these settings, we propose a difficulty-aware attention mechanism to properly handle hard samples or regions by taking structural information into consideration so that we can better deal with the irregular distribution of medical data. Furthermore, we investigate the loss functions of various GANs and propose using the binary cross entropy loss to train the proposed adversarial system so that we can retain the unlimited modeling capacity of the discriminator. Experimental results on clinical and challenge datasets show that our proposed network can achieve state-of-the-art segmentation (synthesis) accuracy. Further analysis also indicates that adversarial confidence learning can both improve the visual perception performance and the quantitative performance.

Texture Feature-Based Classification on Transrectal Ultrasound Image for Prostatic Cancer Detection

Prostate cancer is one of the most common cancers in men. Early detection of prostate cancer is the key to successful treatment. Ultrasound imaging is one of the most suitable methods for the early detection of prostate cancer. Although ultrasound images can show cancer lesions, subjective interpretation is not accurate. Therefore, this paper proposes a transrectal ultrasound image analysis method, aiming at characterizing prostate tissue through image processing to evaluate the possibility of malignant tumours. Firstly, the input image is preprocessed by optical density conversion. Then, local binarization and Gaussian Markov random fields are used to extract texture features, and the linear combination is performed. Finally, the fused texture features are provided to SVM classifier for classification. The method has been applied to data set of 342 transrectal ultrasound images obtained from hospitals with an accuracy of 70.93%, sensitivity of 70.00%, and specificity of 71.74%. The experimental results show that it is possible to distinguish cancerous tissues from noncancerous tissues to some extent.

CDA-Net for Automatic Prostate Segmentation in MR Images

Automatic and accurate prostate segmentation is an essential prerequisite for assisting diagnosis and treatment, such as guiding biopsy procedures and radiation therapy. Therefore, this paper proposes a cascaded dual attention network (CDA-Net) for automatic prostate segmentation in MRI scans. The network includes two stages of RAS-FasterRCNN and RAU-Net. Firstly, RAS-FasterRCNN uses improved FasterRCNN and sequence correlation processing to extract regions of interest (ROI) of organs. This ROI extraction serves as a hard attention mechanism to focus the segmentation of the subsequent network on a certain area. Secondly, the addition of residual convolution block and self-attention mechanism in RAU-Net enables the network to gradually focus on the area where the organ exists while making full use of multiscale features. The algorithm was evaluated on the PROMISE12 and ASPS13 datasets and presents the dice similarity coefficient of 92.88% and 92.65%, respectively, surpassing the state-of-the-art algorithms. In a variety of complex slice images, especially for the base and apex of slice sequences, the algorithm also achieved credible segmentation performance.

MS-Net: Multi-Site Network for Improving Prostate Segmentation With Heterogeneous MRI Data

Automated prostate segmentation in MRI is highly demanded for computer-assisted diagnosis. Recently, a variety of deep learning methods have achieved remarkable progress in this task, usually relying on large amounts of training data. Due to the nature of scarcity for medical images, it is important to effectively aggregate data from multiple sites for robust model training, to alleviate the insufficiency of single-site samples. However, the prostate MRIs from different sites present heterogeneity due to the differences in scanners and imaging protocols, raising challenges for effective ways of aggregating multi-site data for network training. In this paper, we propose a novel multi-site network (MS-Net) for improving prostate segmentation by learning robust representations, leveraging multiple sources of data. To compensate for the inter-site heterogeneity of different MRI datasets, we develop Domain-Specific Batch Normalization layers in the network backbone, enabling the network to estimate statistics and perform feature normalization for each site separately. Considering the difficulty of capturing the shared knowledge from multiple datasets, a novel learning paradigm, i.e., Multi-site-guided Knowledge Transfer, is proposed to enhance the kernels to extract more generic representations from multi-site data. Extensive experiments on three heterogeneous prostate MRI datasets demonstrate that our MS-Net improves the performance across all datasets consistently, and outperforms state-of-the-art methods for multi-site learning.

AutoPath: Image-Specific Inference for 3D Segmentation

In recent years, deep convolutional neural networks (CNNs) has made great achievements in the field of medical image segmentation, among which residual structure plays a significant role in the rapid development of CNN-based segmentation. However, the 3D residual networks inevitably bring a huge computational burden to machines for network inference, thus limiting their usages for many real clinical applications. To tackle this issue, we propose AutoPath, an image-specific inference approach for more efficient 3D segmentations. The proposed AutoPath dynamically selects enabled residual blocks regarding different input images during inference, thus effectively reducing total computation without degrading segmentation performance. To achieve this, a policy network is trained using reinforcement learning, by employing the rewards of using a minimal set of residual blocks and meanwhile maintaining accurate segmentation. Experimental results on liver CT dataset show that our approach not only provides efficient inference procedure but also attains satisfactory segmentation performance.

Generalizing Deep Learning for Medical Image Segmentation to Unseen Domains via Deep Stacked Transformation

Recent advances in deep learning for medical image segmentation demonstrate expert-level accuracy. However, application of these models in clinically realistic environments can result in poor generalization and decreased accuracy, mainly due to the domain shift across different hospitals, scanner vendors, imaging protocols, and patient populations etc. Common transfer learning and domain adaptation techniques are proposed to address this bottleneck. However, these solutions require data (and annotations) from the target domain to retrain the model, and is therefore restrictive in practice for widespread model deployment. Ideally, we wish to have a trained (locked) model that can work uniformly well across unseen domains without further training. In this paper, we propose a deep stacked transformation approach for domain generalization. Specifically, a series of n stacked transformations are applied to each image during network training. The underlying assumption is that the "expected" domain shift for a specific medical imaging modality could be simulated by applying extensive data augmentation on a single source domain, and consequently, a deep model trained on the augmented "big" data (BigAug) could generalize well on unseen domains. We exploit four surprisingly effective, but previously understudied, image-based characteristics for data augmentation to overcome the domain generalization problem. We train and evaluate the BigAug model (with n=9 transformations) on three different 3D segmentation tasks (prostate gland, left atrial, left ventricle) covering two medical imaging modalities (MRI and ultrasound) involving eight publicly available challenge datasets. The results show that when training on relatively small dataset (n = 10~32 volumes, depending on the size of the available datasets) from a single source domain: (i) BigAug models degrade an average of 11%(Dice score change) from source to unseen domain, substantially better than conventional augmentation (degrading 39%) and CycleGAN-based domain adaptation method (degrading 25%), (ii) BigAug is better than "shallower" stacked transforms (i.e. those with fewer transforms) on unseen domains and demonstrates modest improvement to conventional augmentation on the source domain, (iii) after training with BigAug on one source domain, performance on an unseen domain is similar to training a model from scratch on that domain when using the same number of training samples. When training on large datasets (n = 465 volumes) with BigAug, (iv) application to unseen domains reaches the performance of state-of-the-art fully supervised models that are trained and tested on their source domains. These findings establish a strong benchmark for the study of domain generalization in medical imaging, and can be generalized to the design of highly robust deep segmentation models for clinical deployment.

3D APA-Net: 3D Adversarial Pyramid Anisotropic Convolutional Network for Prostate Segmentation in MR Images

Accurate and reliable segmentation of the prostate gland using magnetic resonance (MR) imaging has critical importance for the diagnosis and treatment of prostate diseases, especially prostate cancer. Although many automated segmentation approaches, including those based on deep learning have been proposed, the segmentation performance still has room for improvement due to the large variability in image appearance, imaging interference, and anisotropic spatial resolution. In this paper, we propose the 3D adversarial pyramid anisotropic convolutional deep neural network (3D APA-Net) for prostate segmentation in MR images. This model is composed of a generator (i.e., 3D PA-Net) that performs image segmentation and a discriminator (i.e., a six-layer convolutional neural network) that differentiates between a segmentation result and its corresponding ground truth. The 3D PA-Net has an encoder-decoder architecture, which consists of a 3D ResNet encoder, an anisotropic convolutional decoder, and multi-level pyramid convolutional skip connections. The anisotropic convolutional blocks can exploit the 3D context information of the MR images with anisotropic resolution, the pyramid convolutional blocks address both voxel classification and gland localization issues, and the adversarial training regularizes 3D PA-Net and thus enables it to generate spatially consistent and continuous segmentation results. We evaluated the proposed 3D APA-Net against several state-of-the-art deep learning-based segmentation approaches on two public databases and the hybrid of the two. Our results suggest that the proposed model outperforms the compared approaches on three databases and could be used in a routine clinical workflow.