Dilated convolution network with edge fusion block and directional feature maps for cardiac MRI segmentation

The domain of deep learning stimulates medical image analysis, which is a catalyst of scientific research and an essential element of healthcare. Since semantic segmentation techniques empower image processing and quantitative determination in various applications, designing a dedicated solution is challenging and heavily reliant on input data characteristics and hardware constraints. Cardiac magnetic resonance imaging segmentation provides three essential heart structures, left ventricle (LV) cavity, myocardium (MYO), and right ventricle (RV) cavity. In clinical applications, manual contouring is frequently utilized to perform semantic segmentation. A fully automated cardiac magnetic resonance image (CMRI) segmentation technique is becoming more desirable as deep learning-based frameworks advance. Motivated by the power of U-Net, residual network, and deep supervision, this paper proposes a Deep Residual U-Net to achieve better LV, MYO, and RV segmentation in short-axis CMRI. The model is formed with residual connections and has a similar layout to U-Net. It provides three advantages: First, residual connections make deep network training easier. Second, the network's rich skip connections enable feature propagation, allowing for network creation with lower complexity but more remarkable performance. Third, Deep supervision facilitates the loss computation at every feature dimension except at least two, empowering gradients to be implanted more deeply in the network and improving each layer's training in Deep Residual U-Net. Automated Cardiac Diagnosis Challenge (ACDC) 2017 dataset has been used to assess the segmentation efficiency of proposed model. The proposed approach significantly outperformed all other methods, evidencing its supremacy over recent state-of-the-art U-Net-based techniques.

Cardiac cine magnetic resonance imaging (CMRI) is the reference standard for assessing cardiac structure as well as function. However, CMRI data presents large variations among different centers, vendors, and patients with various cardiovascular diseases. Since typical deep-learning-based segmentation methods are usually trained using a limited number of ground truth annotations, they may not generalize well to unseen MR images, due to the variations between the training and testing data. In this study, we proposed an approach towards building a generalizable deep-learning-based model for cardiac structure segmentations from multi-vendor,multi-center and multi-diseases CMRI data. We used a novel combination of image augmentation and a consistency loss function to improve model robustness to typical variations in CMRI data. The proposed image augmentation strategy leverages un-labeled data by a) using CycleGAN to generate images in different styles and b) exchanging the low-frequency features of images from different vendors. Our model architecture was based on an attention-gated U-Net model that learns to focus on cardiac structures of varying shapes and sizes while suppressing irrelevant regions. The proposed augmentation and consistency training method demonstrated improved performance on CMRI images from new vendors and centers. When evaluated using CMRI data from 4 vendors and 6 clinical center, our method was generally able to produce accurate segmentations of cardiac structures.KeywordsMRI segmentationGeneralizationImage augmentationDeep learning

For a considerable amount of time, medical image processing has been an important topic of research in the field of medicine. The advent of deep learning technology has resulted in the development of revolutionary improvements in this particular sector. Because of its remarkable effectiveness in training enormous amounts of data and performing difficult tasks, deep learning has attracted a lot of attention. It has found extensive application in the analysis, diagnosis, and treatment of medical pictures. This work employs the literature review approach to examine and analyze the existing research conducted by academics on deep learning and medical imaging. It also provides a summary of the utilization of deep learning in medical imaging images and highlights the current development trend in this field.

<abstract> <sec><title>Background</title><p>In clinical diagnostics, magnetic resonance imaging (MRI) technology plays a crucial role in the recognition of cardiac regions, serving as a pivotal tool to assist physicians in diagnosing cardiac diseases. Despite the notable success of convolutional neural networks (CNNs) in cardiac MRI segmentation, it remains a challenge to use existing CNNs-based methods to deal with fuzzy information in cardiac MRI. Therefore, we proposed a novel network architecture named DAFNet to comprehensively address these challenges.</p> </sec> <sec><title>Methods</title><p>The proposed method was used to design a fuzzy convolutional module, which could improve the feature extraction performance of the network by utilizing fuzzy information that was easily ignored in medical images while retaining the advantage of attention mechanism. Then, a multi-scale feature refinement structure was designed in the decoder portion to solve the problem that the decoder structure of the existing network had poor results in obtaining the final segmentation mask. This structure further improved the performance of the network by aggregating segmentation results from multi-scale feature maps. Additionally, we introduced the dynamic convolution theory, which could further increase the pixel segmentation accuracy of the network.</p> </sec> <sec><title>Result</title><p>The effectiveness of DAFNet was extensively validated for three datasets. The results demonstrated that the proposed method achieved DSC metrics of 0.942 and 0.885, and HD metricd of 2.50mm and 3.79mm on the first and second dataset, respectively. The recognition accuracy of left ventricular end-diastolic diameter recognition on the third dataset was 98.42%.</p> </sec> <sec><title>Conclusion</title><p>Compared with the existing CNNs-based methods, the DAFNet achieved state-of-the-art segmentation performance and verified its effectiveness in clinical diagnosis.</p> </sec> </abstract>

Cardiac MRI segmentation plays a crucial role in clinical diagnosis for evaluating personalized cardiac performance parameters. Due to the indistinct boundaries and heterogeneous intensity distributions in the cardiac MRI, most existing methods still suffer from two aspects of challenges: inter-class indistinction and intra-class inconsistency. To tackle these two problems, we propose a novel method to exploit the directional feature maps, which can simultaneously strengthen the differences between classes and the similarities within classes. Specifically, we perform cardiac segmentation and learn a direction field pointing away from the nearest cardiac tissue boundary to each pixel via a direction field (DF) module. Based on the learned direction field, we then propose a feature rectification and fusion (FRF) module to improve the original segmentation features, and obtain the final segmentation. The proposed modules are simple yet effective and can be flexibly added to any existing segmentation network without excessively increasing time and space complexity. We evaluate the proposed method on the 2017 MICCAI Automated Cardiac Diagnosis Challenge (ACDC) dataset and a large-scale self-collected dataset, showing good segmentation performance and robust generalization ability of the proposed method. The code is publicly available at https://github.com/c-feng/DirectionalFeature.

Deep learning plays a significant role in transforming medical imaging for disease diagnosis. It uses advanced algorithms, especially Convolutional Neural Networks (CNNs), to automatically learn and extract important features from medical images. This technology helps in detecting, classifying, and diagnosing various diseases, such as different types of cancer, brain disorders like aneurysms and strokes, heart diseases, and respiratory conditions. Deep learning improves the accuracy and efficiency of diagnostic workflows and reduces the workload for healthcare professionals. Despite its many advantages, deep learning faces challenges related to data availability, model interpretability, and clinical validation. This review highlights the current applications, performance evaluation methods, and challenges of deep learning in medical imaging for disease diagnosis.

Internal images of the body may be captured using X-ray technology, which helps doctors diagnose and treat a wide range of medical issues. Recent studies have focused on optimizing the usage of deep learning algorithms to boost medical imaging's accuracy and productivity. Deep learning AI makes use of massive datasets to train computer algorithms to recognize patterns, which are then used to make predictions or classifications about fresh data. This method has shown promise in facilitating quicker and more accurate detection of conditions including lung cancer and bone fractures using X-ray images. Patient outcomes may improve with the use of deep learning algorithms in medical imaging. Healthcare systems may be strengthened, and the effects of health crises lessened by facilitating faster and more accurate diagnosis. Investing in initiatives that promote the use of deep learning in medical imaging will help get us there. A bespoke deep learning model was able to classify 70% of a dataset evenly between the healthy, COVID, and pneumonia patients, yielding an astonishing 97.96% accuracy. Applying deep learning algorithms to the analysis of X-ray images has shown great potential for improving the efficiency and precision of medical diagnoses. Better health outcomes may result from this enhanced ability to identify abnormalities and illnesses such as lung cancer and bone fractures. The healthcare industry stands to benefit greatly from adopting programs that encourage the integration of deep learning in medical imaging and might even save lives by doing so.

Along with the development of medical imaging technology, medical image analysis has become a crucial part of clinical diagnosis. However, traditional image analysis methods rely on manual feature extraction and expert evaluation, triggering problems such as low efficiency and lack of accuracy. Deep learning techniques, especially convolutional neural networks (CNNs), have made significant breakthroughs in medical image analysis in recent years. The aim of this paper is to study the application of deep learning in medical imaging, and explore the advantages and development potential in early disease diagnosis, image segmentation and feature extraction. This paper reviews the basic principles of deep learning and analyses the applicability of deep learning in medical imaging by combining the characteristics of medical imaging. Through systematic analysis of existing literature, this paper summarises the remarkable achievements of deep learning in processing, analysing and applying it to various medical images, such as Optical Coherence Tomography (OCT), Magnetic Resonance Imaging (MRI) and so on. It has been shown that deep learning methods can significantly improve the accuracy and efficiency of medical image processing and demonstrate performance due to traditional methods in tasks such as classification. However, how to improve the interpretability of models and the generalisability of clinical applications are still current research hotspots and challenges.

The use of machine learning (ML) has been increasing rapidly in the medical imaging field, including computer-aided diagnosis (CAD), radiomics, and medical image analysis. Recently, an ML area called deep learning emerged in the computer vision field and became very popular in many fields. It started from an event in late 2012, when a deep-learning approach based on a convolutional neural network (CNN) won an overwhelming victory in the best-known worldwide computer vision competition, ImageNet Classification. Since then, researchers in virtually all fields, including medical imaging, have started actively participating in the explosively growing field of deep learning. In this paper, the area of deep learning in medical imaging is overviewed, including (1) what was changed in machine learning before and after the introduction of deep learning, (2) what is the source of the power of deep learning, (3) two major deep-learning models: a massive-training artificial neural network (MTANN) and a convolutional neural network (CNN), (4) similarities and differences between the two models, and (5) their applications to medical imaging. This review shows that ML with feature input (or feature-based ML) was dominant before the introduction of deep learning, and that the major and essential difference between ML before and after deep learning is the learning of image data directly without object segmentation or feature extraction; thus, it is the source of the power of deep learning, although the depth of the model is an important attribute. The class of ML with image input (or image-based ML) including deep learning has a long history, but recently gained popularity due to the use of the new terminology, deep learning. There are two major models in this class of ML in medical imaging, MTANN and CNN, which have similarities as well as several differences. In our experience, MTANNs were substantially more efficient in their development, had a higher performance, and required a lesser number of training cases than did CNNs. "Deep learning", or ML with image input, in medical imaging is an explosively growing, promising field. It is expected that ML with image input will be the mainstream area in the field of medical imaging in the next few decades.

Delineation of right ventricular cavity (RVC), left ventricular myocardium (LVM) and left ventricular cavity (LVC) are common tasks in the clinical diagnosis of cardiac related diseases, especially in the basis of advanced magnetic resonance imaging (MRI) techniques. Recently, despite deep learning techniques being widely employed in solving segmentation tasks in a variety of medical images, the sheer volume and complexity of the data in some applications such as cine cardiac MRI pose significant challenges for the accurate and efficient segmentation. In cine cardiac MRI we need to segment both short and long axis 2D images. In this paper, we focus on the automated segmentation of short-axis cardiac MRI images. We first introduce the deep layer aggregation (DLA) method to augment the standard deep learning architecture with deeper aggregation to better fuse information across layers, which is particularly suitable for the cardiac MRI segmentation, due to the complexity of the cardiac boundaries appearance and acquisition resolution during a cardiac cycle. In our solution, we develop a modified DLA framework by embedding Refinement Residual Block (RRB) and Channel Attention Block (CAB). Experimental results validate the superior performance of our proposed method for the cardiac structures segmentation in comparison with state-of-the-art. Moreover, we demonstrate its potential use case in the quantitative analysis of cardiac dyssynchrony.

The recent development of a real-time magnetic resonance imaging (MRI) technique with 20 to 30 ms temporal resolution allows for imaging multiple consecutive heart cycles, without the need for breath holding or ECG synchronization. Manual analysis of the resulting image series is no longer feasible because of their length. We propose a region-based algorithm for automatically segmenting the myocardium in consecutive heart cycles based on local context and prior knowledge. The method was evaluated on ten real-time MRI series and compared to segmentations by two observers, with promising results. We show that our approach enables a multicycle analysis of the heart function robust to breathing and arrhythmia.

With the development of the city, the traffic crowd congested scene is increasing frequency. And the traffic crowd congested may bring disaster. It is important for city traffic management to recognize traffic crowd congested scene. However, the traffic crowd scene is dynamically and the visual scales are varied. Due to the multi-scale problem, it is hard to distinguish the congested traffic crowd scene. To solve the multiple scales problem in traffic crowd congested scene recognition, in this paper, a traffic crowd congested scene recognition method based on dilated convolution network is proposed, which combines the dilated convolution and VGG16 network for traffic crowd congested scene recognition. To verify the proposed method, the experiments are implemented on two crowd datasets including the CUHK Crowd dataset and Normal-abnormal Crowd dataset. And the experimental results are compared with three states of the art methods. The experimental results demonstrate that the performance of the proposed method is more effective in congested traffic crowd scene recognition. Compared with the three state of the art methods, the average accuracy value, and the average AUC values of the proposed method are improved by 15.87\(\%\) and 11.58\(\%\) respectively.KeywordsTraffic crowdCongested sceneDilated convolutionDeep learningTwo-stream convolution

The success of deep learning in medical imaging is mostly achieved at the cost of a large labeled data set. Semi-supervised learning (SSL) provides a promising solution by leveraging the structure of unlabeled data to improve learning from a small set of labeled data. Self-ensembling is a simple approach used in SSL to encourage consensus among ensemble predictions of unknown labels, improving generalization of the model by making it more insensitive to the latent space. Currently, such an ensemble is obtained by randomization such as dropout regularization and random data augmentation. In this work, we hypothesize – from the generalization perspective – that self-ensembling can be improved by exploiting the stochasticity of a disentangled latent space. To this end, we present a stacked SSL model that utilizes unsupervised disentangled representation learning as the stochastic embedding for self-ensembling. We evaluate the presented model for multi-label classification using chest X-ray images, demonstrating its improved performance over related SSL models as well as the interpretability of its disentangled representations.

The use of deep learning in medical imaging to improve spine care: A scoping review of current literature and clinical applications

Coronary CT angiography (CCTA) is the only non-invasive imaging technique that reliably depicts the anatomic extent of Coronary Artery Disease (CAD). While occlusion remains a highly predictive indicator of major cardiovascular events (MACE), there is growing evidence that the presence and characteristics of coronary atherosclerosis provide additional prognostic information. In CCTA calcified plaques display high-intensity Hounsfield Units (HU) representative features while more complex representations characterize high-risk soft plaques. As such, accurate identification and quantification is burdensome and time consuming because of the limited temporal, spatial and contrast resolutions of X-ray scanners. Despite the success of deep learning in medical imaging, automatic localization of coronary plaques and especially soft plaques remains a challenging subject in CCTA vessel analysis. For this study, 150 CCTA scans were retrospectively collected. All patients were accepted at triage with minimal to severe CAD suspicion. Selection was carried out with uniform CAD-RADS severity distribution which normally follows an exponential decay function, thus obtaining a higher than normal concentration of plaques. The proposed method outperforms the state of the art for the localization of diverse types of plaques by exploiting the self-attention mechanism of transformers networks to embed contextual features of the coronary tree.