Search

Glomerulus morphology on renal pathology images provides valuable diagnosis and outcome prediction information. To provide better care, an efficient, standardized, and scalable method is urgently needed to optimize the time-consuming and labor-intensive interpretation process by renal pathologists. This paper proposes a deep convolutional neural network (CNN)-based approach to automatically detect and classify glomeruli with different stains in renal pathology images. In the glomerulus detection stage, this paper proposes a flattened Xception with a feature pyramid network (FX-FPN). The FX-FPN is employed as a backbone in the framework of faster region-based CNN to improve glomerulus detection performance. In the classification stage, this paper considers classifications of five glomerulus morphologies using a flattened Xception classifier. To endow the classifier with higher discriminability, this paper proposes a generative data augmentation approach for patch-based glomerulus morphology augmentation. New glomerulus patches of different morphologies are generated for data augmentation through the cycle-consistent generative adversarial network (CycleGAN). The single detection model shows the F 1 score up to 0.9524 in H&E and PAS stains. The classification result shows that the average sensitivity and specificity are 0.7077 and 0.9316, respectively, by using the flattened Xception with the original training data. The sensitivity and specificity increase to 0.7623 and 0.9443, respectively, by using the generative data augmentation. Comparisons with different deep CNN models show the effectiveness and superiority of the proposed approach. • Collection of renal pathological images with annotations of glomerulus classes. • A flattened Xception with feature pyramid network (FX-FPN) for glomerulus detection. • Generative glomerulus morphology augmentation to improve classification performance. • A two-stage detection and classification framework with performance evaluation. [ABSTRACT FROM AUTHOR]

Gastric precancerous lesions (GPL) significantly elevate the risk of gastric cancer, and precise diagnosis and timely intervention are critical for patient survival. Due to the elusive pathological features of precancerous lesions, the early detection rate is less than 10%, which hinders lesion localization and diagnosis. In this paper, we provide a GPL pathological dataset and propose a novel method for improving the segmentation accuracy on a limited-scale dataset, namely RGB and Hyperspectral dual-modal pathological image Cross-attention U-Net (CrossU-Net). Specifically, we present a self-supervised pre-training model for hyperspectral images to serve downstream segmentation tasks. Secondly, we design a dual-stream U-Net-based network to extract features from different modal images. To promote information exchange between spatial information in RGB images and spectral information in hyperspectral images, we customize the cross-attention mechanism between the two networks. Furthermore, we use an intermediate agent in this mechanism to improve computational efficiency. Finally, we add a distillation loss to align predicted results for both branches, improving network generalization. Experimental results show that our CrossU-Net achieves accuracy and Dice of 96.53% and 91.62%, respectively, for GPL lesion segmentation, providing a promising spectral research approach for the localization and subsequent quantitative analysis of pathological features in early diagnosis. • This paper presented a dual-modality CrossU-Net for lesion segmentation of gastric precancerous from RGB and HSI images. • Developed a Spectral-wise Siamese (SpeSiam) self-supervised pre-training module that focuses on spectral analysis. • Designed a dual-stream Cross-attention U-Net network to improve lesion tissue identification. • Demonstrated the effectiveness of our approach through comparisons with single-modal methods for pathological diagnosis. [ABSTRACT FROM AUTHOR]

In this paper, we hypothesize that it is possible to localize image regions of preclinical tumors in a Chest X-ray (CXR) image by a weakly-supervised training of a survival prediction model using a dataset containing CXR images of healthy patients and their time-to-death label. These visual explanations can empower clinicians in early lung cancer detection and increase patient awareness of their susceptibility to the disease. To test this hypothesis, we train a censor-aware multi-class survival prediction deep learning classifier that is robust to imbalanced training, where classes represent quantized number of days for time-to-death prediction. Such multi-class model allows us to use post-hoc interpretability methods, such as Grad-CAM, to localize image regions of preclinical tumors. For the experiments, we propose a new benchmark based on the National Lung Cancer Screening Trial (NLST) dataset to test weakly-supervised preclinical tumor localization and survival prediction models, and results suggest that our proposed method shows state-of-the-art C-index survival prediction and weakly-supervised preclinical tumor localization results. To our knowledge, this constitutes a pioneer approach in the field that is able to produce visual explanations of preclinical events associated with survival prediction results. • First weakly-supervised survival prediction method to localize preclinical tumors. • State-of-the-art survival time prediction results on the NLST dataset. • Innovative explainable censor-aware multi-class survival prediction. • Innovative survival prediction robust to imbalanced time-to-death distribution. • New benchmark for explainable preclinical survival prediction using the NLST dataset. [ABSTRACT FROM AUTHOR]

We address the problem of lung CT image registration, which underpins various diagnoses and treatments for lung diseases. The main crux of the problem is the large deformation that the lungs undergo during respiration. This physiological process imposes several challenges from a learning point of view. In this paper, we propose a novel training scheme, called stochastic decomposition, which enables deep networks to effectively learn such a difficult deformation field during lung CT image registration. The key idea is to stochastically decompose the deformation field, and supervise the registration by synthetic data that have the corresponding appearance discrepancy. The stochastic decomposition allows for revealing all possible decompositions of the deformation field. At the learning level, these decompositions can be seen as a prior to reduce the ill-posedness of the registration yielding to boost the performance. We demonstrate the effectiveness of our framework on Lung CT data. We show, through extensive numerical and visual results, that our technique outperforms existing methods. • A novel and effective training algorithm for deformable lung CT registration. • Stochastic decomposition of large deformation and supervision with surrogate images. • The surrogate image has corresponding intensity discrepancy at decomposition phase. [ABSTRACT FROM AUTHOR]

Automated Motion Artefact Detection (MAD) in Magnetic Resonance Imaging (MRI) is a field of study that aims to automatically flag motion artefacts in order to prevent the requirement for a repeat scan. In this paper, we identify and tackle the three current challenges in the field of automated MAD; (1) reliance on fully-supervised training, meaning they require specific examples of Motion Artefacts (MA), (2) inconsistent use of benchmark datasets across different works and use of private datasets for testing and training of newly proposed MAD techniques and (3) a lack of sufficiently large datasets for MRI MAD. To address these challenges, we demonstrate how MAs can be identified by formulating the problem as an unsupervised Anomaly Detection (AD) task. We compare the performance of three State-of-the-Art AD algorithms DeepSVDD, Interpolated Gaussian Descriptor and FewSOME on two open-source Brain MRI datasets on the task of MAD and MA severity classification, with FewSOME achieving a MAD AUC > 90 % on both datasets and a Spearman Rank Correlation Coefficient of 0.8 on the task of MA severity classification. These models are trained in the few shot setting, meaning large Brain MRI datasets are not required to build robust MAD algorithms. This work also sets a standard protocol for testing MAD algorithms on open-source benchmark datasets. In addition to addressing these challenges, we demonstrate how our proposed 'anomaly-aware' scoring function improves FewSOME's MAD performance in the setting where one and two shots of the anomalous class are available for training. Code available at https://github.com/niamhbelton/Unsupervised-Brain-MRI-Motion-Artefact-Detection/. • This work investigates unsupervised approaches to MAD and MA severity classification. • As the models are unsupervised, the requirement to acquire MRIs with MAs is eliminated. • FewSOME achieves AUCs > 0.9 on two datasets having trained on a small dataset of 30 MRIs. • This work shows how the anomaly scores correlate with MA severity. • This work proposes an 'anomaly-aware' scoring function. • This work provides a standardised protocol that can be used as a benchmark for future MAD analyses. [ABSTRACT FROM AUTHOR]

Colonoscopy is the choice procedure to diagnose, screening, and treat the colon and rectum cancer, from early detection of small precancerous lesions (polyps), to confirmation of malign masses. However, the high variability of the organ appearance and the complex shape of both the colon wall and structures of interest make this exploration difficult. Learned visuospatial and perceptual abilities mitigate technical limitations in clinical practice by proper estimation of the intestinal depth. This work introduces a novel methodology to estimate colon depth maps in single frames from monocular colonoscopy videos. The generated depth map is inferred from the shading variation of the colon wall with respect to the light source, as learned from a realistic synthetic database. Briefly, a classic convolutional neural network architecture is trained from scratch to estimate the depth map, improving sharp depth estimations in haustral folds and polyps by a custom loss function that minimizes the estimation error in edges and curvatures. The network was trained by a custom synthetic colonoscopy database herein constructed and released, composed of 248 400 frames (47 videos), with depth annotations at the level of pixels. This collection comprehends 5 subsets of videos with progressively higher levels of visual complexity. Evaluation of the depth estimation with the synthetic database reached a threshold accuracy of 95.65%, and a mean-RMSE of 0. 451 cm , while a qualitative assessment with a real database showed consistent depth estimations, visually evaluated by the expert gastroenterologist coauthoring this paper. Finally, the method achieved competitive performance with respect to another state-of-the-art method using a public synthetic database and comparable results in a set of images with other five state-of-the-art methods. Additionally, three-dimensional reconstructions demonstrated useful approximations of the gastrointestinal tract geometry. Code for reproducing the reported results and the dataset are available at https://github.com/Cimalab-unal/ColonDepthEstimation. [Display omitted] • A convolutional neural network, the SfSNet, estimates a pixel-wise depth map of the colon in a raw frame from the colonoscopy video. • A realistic synthetic colonoscopy database with depth annotations that is publicly available, with progressively higher levels of visual complexity. • A custom loss function which drives the network by minimizing the estimation error at folds and polyps. • Two training strategies are evaluated to optimize the depth estimation in colon folds and precancerous polyps. [ABSTRACT FROM AUTHOR]

Rib cross-sectional shapes (characterized by the outer contour and cortical bone thickness) affect the rib mechanical response under impact loading, thereby influence the rib injury pattern and risk. A statistical description of the rib shapes or their correlations to anthropometrics is a prerequisite to the development of numerical human body models representing target demographics. Variational autoencoders (VAE) as anatomical shape generators remain to be explored in terms of utilizing the latent vectors to control or interpret the representativeness of the generated results. In this paper, we propose a pipeline for developing a multi-rib cross-sectional shape generative model from CT images, which consists of the achievement of rib cross-sectional shape data from CT images using an anatomical indexing system and regular grids, and a unified framework to fit shape distributions and associate shapes to anthropometrics for different rib categories. Specifically, we collected CT images including 3193 ribs, surface regular grid is generated for each rib based on anatomical coordinates, the rib cross-sectional shapes are characterized by nodal coordinates and cortical bone thickness. The tensor structure of shape data based on regular grids enable the implementation of CNNs in the conditional variational autoencoder (CVAE). The CVAE is trained against an auxiliary classifier to decouple the low-dimensional representations of the inter- and intra- variations and fit each intra-variation by a Gaussian distribution simultaneously. Random tree regressors are further leveraged to associate each continuous intra-class space with the corresponding anthropometrics of the subjects, i.e., age, height and weight. As a result, with the rib class labels and the latent vectors sampled from Gaussian distributions or predicted from anthropometrics as the inputs, the decoder can generate valid rib cross-sectional shapes of given class labels (male/female, 2nd to 11th ribs) for arbitrary populational percentiles or specific age, height and weight, which paves the road for future biomedical and biomechanical studies considering the diversity of rib shapes across the population. • A deep learning-based pipeline is proposed for developing multi-rib shape generative model. • Rib cross-sectional shape data are achieved from CT images using an indexing system and regular grids. • A unified framework is proposed to fit shape distributions and associate shapes to anthropometrics for different ribs. • Valid rib cross-sectional shapes can be generated for arbitrary population percentiles or anthropometrics. [ABSTRACT FROM AUTHOR]

Vascular structure segmentation in intravascular ultrasound (IVUS) images plays an important role in pre-procedural evaluation of percutaneous coronary intervention (PCI). However, vascular structure segmentation in IVUS images has the challenge of structure-dependent distractions. Structure-dependent distractions are categorized into two cases, structural intrinsic distractions and inter-structural distractions. Traditional machine learning methods often rely solely on low-level features, overlooking high-level features. This way limits the generalization of these methods. The existing semantic segmentation methods integrate low-level and high-level features to enhance generalization performance. But these methods also introduce additional interference, which is harmful to solving structural intrinsic distractions. Distraction cue methods attempt to address structural intrinsic distractions by removing interference from the features through a unique decoder. However, they tend to overlook the problem of inter-structural distractions. In this paper, we propose distraction-aware hierarchical learning (DHL) for vascular structure segmentation in IVUS images. Inspired by distraction cue methods for removing interference in a decoder, the DHL is designed as a hierarchical decoder that gradually removes structure-dependent distractions. The DHL includes global perception process, distraction perception process and structural perception process. The global perception process and distraction perception process remove structural intrinsic distractions then the structural perception process removes inter-structural distractions. In the global perception process, the DHL searches for the coarse structural region of the vascular structures on the slice of IVUS sequence. In the distraction perception process, the DHL progressively refines the coarse structural region of the vascular structures to remove structural distractions. In the structural perception process, the DHL detects regions of inter-structural distractions in fused structure features then separates them. Extensive experiments on 361 subjects show that the DHL is effective (e.g., the average Dice is greater than 0.95), and superior to ten state-of-the-art IVUS vascular structure segmentation methods. • We propose the distraction-aware hierarchical learning (DHL) for vascular structure segmentation in IVUS images. • We propose new structural perception module to detect inconspicuous regions from fused structure features. • We explore a strategy of structural grouping to improve the efficiency of the refinement of structures. • We use new trajectory attention to solve global morphological variability. • The DHL explores for the first time the segmentation of IVUS myocardial bridges. [ABSTRACT FROM AUTHOR]

The absence of bone wall located in the jugular bulb and sigmoid sinus of the temporal bone is one of the important reasons for pulsatile tinnitus. Automatic and accurate detection of these abnormal singes in CT slices has important theoretical significance and clinical value. Due to the shortage of abnormal samples, imbalanced samples, small inter-class differences, and low interpretability, existing deep-learning methods are greatly challenged. In this paper, we proposed a sub-features orthogonal decoupling model, which can effectively disentangle the representation features into class-specific sub-features and class-independent sub-features in a latent space. The former contains the discriminative information, while, the latter preserves information for image reconstruction. In addition, the proposed method can generate image samples using category conversion by combining the different class-specific sub-features and the class-independent sub-features, achieving corresponding mapping between deep features and images of specific classes. The proposed model improves the interpretability of the deep model and provides image synthesis methods for downstream tasks. The effectiveness of the method was verified in the detection of bone wall absence in the temporal bone jugular bulb and sigmoid sinus. • A bone wall absence anomaly detection network in temporal bone CT is presented. • A sub-feature orthogonal model is designed to decouple the discriminative features. • Experimental results verified the effectiveness of the bone wall absence detection. [ABSTRACT FROM AUTHOR]

Medical images play a vital role in medical analysis by providing crucial information about patients' pathological conditions. However, the quality of these images can be compromised by many factors, such as limited resolution of the instruments, artifacts caused by movements, and the complexity of the scanned areas. As a result, low-resolution (LR) images cannot provide sufficient information for diagnosis. To address this issue, researchers have attempted to apply image super-resolution (SR) techniques to restore the high-resolution (HR) images from their LR counterparts. However, these techniques are designed for generic images, and thus suffer from many challenges unique to medical images. An obvious one is the diversity of the scanned objects; for example, the organs, tissues, and vessels typically appear in different sizes and shapes, and are thus hard to restore with standard convolution neural networks (CNNs). In this paper, we develop a dynamic-local learning framework to capture the details of these diverse areas, consisting of deformable convolutions with adjustable kernel shapes. Moreover, the global information between the tissues and organs is vital for medical diagnosis. To preserve global information, we propose pixel–pixel and patch–patch global learning using a non-local mechanism and a vision transformer (ViT), respectively. The result is a novel CNN-ViT neural network with Local-to-Global feature learning for medical image SR, referred to as LGSR, which can accurately restore both local details and global information. We evaluate our method on six public datasets and one large-scale private dataset, which include five different types of medical images (i.e. , Ultrasound, OCT, Endoscope, CT, and MRI images). Experiments show that the proposed method achieves superior PSNR/SSIM and visual performance than the state of the arts with competitive computational costs, measured in network parameters, runtime, and FLOPs. What is more, the experiment conducted on OCT image segmentation for the downstream task demonstrates a significantly positive performance effect of LGSR. • We proposed a novel CNN-ViT model, LGSR, for multi-modality medical image SR. • We defined a local-to-global learning framework to restore information thoroughly. • LGSR costs fewer parameters than CNN SOTAs, and is faster than CNN-ViT SOTAs. [ABSTRACT FROM AUTHOR]

Knee OsteoArthritis (OA) is a prevalent chronic condition, affecting a significant proportion of the global population. Detecting knee OA is crucial as the degeneration of the knee joint is irreversible. In this paper, we introduce a semi-supervised multi-view framework and a 3D CNN model for detecting knee OA using 3D Magnetic Resonance Imaging (MRI) scans. We introduce a semi-supervised learning approach combining labeled and unlabeled data to improve the performance and generalizability of the proposed model. Experimental results show the efficacy of our proposed approach in detecting knee OA from 3D MRI scans using a large cohort of 4297 subjects. An ablation study was conducted to investigate the contributions of various components of the proposed model, providing insights into the optimal design of the model. Our results indicate the potential of the proposed approach to improve the accuracy and efficiency of OA diagnosis. The proposed framework reported an AUC of 93.20% for the detection of knee OA. • A novel deep learning model is introduced for knee OA detection from MR images. • A semi-supervised approach is proposed to increase the training dataset. • A multiview framework using multiple MRI views is proposed for knee OA detection. • Ablation studies and Grad-CAM heatmaps show the potential of the proposed approach. • Experiments were achieved on the public OsteoArthritis Initiative (OAI) database. [ABSTRACT FROM AUTHOR]

Bipolar disorder (BD) is characterized by recurrent episodes of depression and mild mania. In this paper, to address the common issue of insufficient accuracy in existing methods and meet the requirements of clinical diagnosis, we propose a framework called Spatio-temporal Feature Fusion Transformer (STF2Former). It improves on our previous work — MFFormer by introducing a Spatio-temporal Feature Aggregation Module (STFAM) to learn the temporal and spatial features of rs-fMRI data. It promotes intra-modality attention and information fusion across different modalities. Specifically, this method decouples the temporal and spatial dimensions and designs two feature extraction modules for extracting temporal and spatial information separately. Extensive experiments demonstrate the effectiveness of our proposed STFAM in extracting features from rs-fMRI, and prove that our STF2Former can significantly outperform MFFormer and achieve much better results among other state-of-the-art methods. [Display omitted] An overview of the proposed method Spatio-temporal Feature Fusion Transformer (STF2Former). Resting-state functional magnetic resonance imaging (rs-fMRI) data undergoes feature extraction through 2D convolution layers with 1x3 and 3x1 kernels to extract temporal and spatial features. Residual structures are used to preserve useful information. T1w goes through five layers of 3D convolution and max pooling to obtain feature representation. In the final feature extraction stage, a transformer self-attention mechanism is introduced to learn intermediate representations that fuse across multiple modalities. After a series of feature extraction, the feature maps are concatenated and fed into two convolutional layers and an MLP, producing logits as the output. • We propose a new framework which significantly improves the performance compared to MFFormer. • We introduce a scheme to utilize the temporal and spatial features of rs-fMRI data fully. • Our results demonstrate that STF2Former achieves SOTA performance in BD disease diagnosis tasks. [ABSTRACT FROM AUTHOR]

Vascular landmark detection plays an important role in medical analysis and clinical treatment. However, due to the complex topology and similar local appearance around landmarks, the popular heatmap regression based methods always suffer from the landmark confusion problem. Vascular landmarks are connected by vascular segments and have special spatial correlations, which can be utilized for performance improvement. In this paper, we propose a multi-task global optimization-based framework for accurate and automatic vascular landmark detection. A multi-task deep learning network is exploited to accomplish landmark heatmap regression, vascular semantic segmentation, and orientation field regression simultaneously. The two auxiliary objectives are highly correlated with the heatmap regression task and help the network incorporate the structural prior knowledge. During inference, instead of performing a max-voting strategy, we propose a global optimization-based post-processing method for final landmark decision. The spatial relationships between neighboring landmarks are utilized explicitly to tackle the landmark confusion problem. We evaluated our method on a cerebral MRA dataset with 564 volumes, a cerebral CTA dataset with 510 volumes, and an aorta CTA dataset with 50 volumes. The experiments demonstrate that the proposed method is effective for vascular landmark localization and achieves state-of-the-art performance. • End-to-end deep-learning based method for vascular landmark detection. • Multi-task network to incorporate structural prior information. • Global optimization-based post-processing algorithm for landmark decision. • A new evaluation metric to evaluate topological errors in landmark detection task. • Experiments on datasets with different vascular structures and imaging modalities. [ABSTRACT FROM AUTHOR]

Brain tumors have become a severe medical complication in recent years due to their high fatality rate. Radiologists segment the tumor manually, which is time-consuming, error-prone, and expensive. In recent years, automated segmentation based on deep learning has demonstrated promising results in solving computer vision problems such as image classification and segmentation. Brain tumor segmentation has recently become a prevalent task in medical imaging to determine the tumor location, size, and shape using automated methods. Many researchers have worked on various machine and deep learning approaches to determine the most optimal solution using the convolutional methodology. In this review paper, we discuss the most effective segmentation techniques based on the datasets that are widely used and publicly available. We also proposed a survey of federated learning methodologies to enhance global segmentation performance and ensure privacy. A comprehensive literature review is suggested after studying more than 100 papers to generalize the most recent techniques in segmentation and multi-modality information. Finally, we concentrated on unsolved problems in brain tumor segmentation and a client-based federated model training strategy. Based on this review, future researchers will understand the optimal solution path to solve these issues. • Presented effective segmentation techniques following publicly available datasets. • Surveyed many papers to generalize modality and multi-modality information systems. • Surveyed federated learning methods to enhance segmentation and privacy assurance. • Identified problems in brain tumor segmentation and client-based federated model. [ABSTRACT FROM AUTHOR]

Lung granuloma is a very common lung disease, and its specific diagnosis is important for determining the exact cause of the disease as well as the prognosis of the patient. And, an effective lung granuloma detection model based on computer-aided diagnostics (CAD) can help pathologists to localize granulomas, thereby improving the efficiency of the specific diagnosis. However, for lung granuloma detection models based on CAD, the significant size differences between granulomas and how to better utilize the morphological features of granulomas are both critical challenges to be addressed. In this paper, we propose an automatic method CRDet to localize granulomas in histopathological images and deal with these challenges. We first introduce the multi-scale feature extraction network with self-attention to extract features at different scales at the same time. Then, the features will be converted to circle representations of granulomas by circle representation detection heads to achieve the alignment of features and ground truth. In this way, we can also more effectively use the circular morphological features of granulomas. Finally, we propose a center point calibration method at the inference stage to further optimize the circle representation. For model evaluation, we built a lung granuloma circle representation dataset named LGCR, including 288 images from 50 subjects. Our method yielded 0.316 m A P and 0.571 m A R , outperforming the state-of-the-art object detection methods on our proposed LGCR. • Proposing a circle representation detector to effectively identify lung granulomas. • Using a multi-scale feature extraction network for bidirectional feature learning. • Designing center point calibration to enhance interpretability and generalization. [ABSTRACT FROM AUTHOR]

Medical image reports are integral to clinical decision-making and patient management. Despite their importance, the confidentiality and private nature of medical data pose significant issues for the sharing and analysis of medical image data. This paper addresses these concerns by introducing a multimodal federated learning-based methodology for medical image reporting. This methodology harnesses distributed computing for co-training models across various medical institutions. Under the federated learning framework, every medical institution is capable of training the model locally and aggregating the updated model parameters to curate a top-tier medical image report model. Initially, we advocate for an architecture facilitating multimodal federated learning, including model creation, parameter consolidation, and algorithm enhancement steps. In the model selection phase, we introduce a deep learning-based strategy that utilizes multimodal data for training to produce medical image reports. In the parameter aggregation phase, the federal average algorithm is applied to amalgamate model parameters trained by each institution, which leads to a comprehensive global model. In addition, we introduce an evidence-based optimization algorithm built upon the federal average algorithm. The efficacy of the proposed architecture and scheme is showcased through a series of experiments. Our experimental results validate the proficiency of the proposed multimodal federated learning approach in generating medical image reports. Compared to conventional centralized learning methods, our proposal not only enhances the protection of patient confidentiality but also enriches the accuracy and overall quality of medical image reports. Through this research, we offer a novel solution for the privacy issues linked with the sharing and analyzing of medical data. Expected to assume a crucial role in medical image report generation and other medical applications, the multimodal federated learning method is set to deliver more precise, efficient, and privacy-secured medical services for healthcare professionals and patients. • A generative model tailored for multimodal joint learning. • Proposed a new algorithm based on the traditional FedAvg and achieved an overall improvement. • Using federated learning to improve the quality of medical image reports. [ABSTRACT FROM AUTHOR]

Automatic retinal arteriovenous classification can assist ophthalmologists in disease early diagnosis. Deep learning-based methods and topological graph-based methods have become the main solutions for retinal arteriovenous classification in recent years. This paper reviews the automatic retinal arteriovenous classification methods from 2003 to 2022. Firstly, we compare different methods and provide comparison tables of the summary results. Secondly, we complete the classification of the public arteriovenous classification datasets and provide the annotation development tables of different datasets. Finally, we sort out the challenges of evaluation methods and provide a comprehensive evaluation system. Quantitative and qualitative analysis shows the changes in research hotspots over time, Quantitative and qualitative analyses reveal the evolution of research hotspots over time, highlighting the significance of exploring the integration of deep learning with topological information in future research. • Summarize different methods in automatic retinal arteriovenous classification. • Compile the public datasets and their annotation improvement processes. • Summarize the flaws of existing evaluation systems and propose improvements. • Identify research future trends through qualitative and quantitative analysis. [ABSTRACT FROM AUTHOR]

Automated medical image segmentation plays a crucial role in diverse clinical applications. The high annotation costs of fully-supervised medical segmentation methods have spurred a growing interest in semi-supervised methods. Existing semi-supervised medical segmentation methods train the teacher segmentation network using labeled data to establish pseudo labels for unlabeled data. The quality of these pseudo labels is constrained as these methods fail to effectively address the significant bias in the data distribution learned from the limited labeled data. To address these challenges, this paper introduces an innovative Correspondence-based Generative Bayesian Deep Learning (C-GBDL) model. Built upon the teacher–student architecture, we design a multi-scale semantic correspondence method to aid the teacher model in generating high-quality pseudo labels. Specifically, our teacher model, embedded with the multi-scale semantic correspondence, learns a better-generalized data distribution from input volumes by feature matching with the reference volumes. Additionally, a double uncertainty estimation schema is proposed to further rectify the noisy pseudo labels. The double uncertainty estimation takes the predictive entropy as the first uncertainty estimation and takes the structural similarity between the input volume and its corresponding reference volumes as the second uncertainty estimation. Four groups of comparative experiments conducted on two public medical datasets demonstrate the effectiveness and the superior performance of our proposed model. Our code is available on https://github.com/yumjoo/C-GBDL. • A C-GBDL model is proposed to capture data distributions with better generalization. • The double uncertainty estimation scheme is designed to refine the pseudo labels. • Comparative experiments have demonstrated the superiority of our methods. [ABSTRACT FROM AUTHOR]

Low resolution of positron emission tomography (PET) limits its diagnostic performance. Deep learning has been successfully applied to achieve super-resolution PET. However, commonly used supervised learning methods in this context require many pairs of low- and high-resolution (LR and HR) PET images. Although unsupervised learning utilizes unpaired images, the results are not as good as that obtained with supervised deep learning. In this paper, we propose a quasi-supervised learning method, which is a new type of weakly-supervised learning methods, to recover HR PET images from LR counterparts by leveraging similarity between unpaired LR and HR image patches. Specifically, LR image patches are taken from a patient as inputs, while the most similar HR patches from other patients are found as labels. The similarity between the matched HR and LR patches serves as a prior for network construction. Our proposed method can be implemented by designing a new network or modifying an existing network. As an example in this study, we have modified the cycle-consistent generative adversarial network (CycleGAN) for super-resolution PET. Our numerical and experimental results qualitatively and quantitatively show the merits of our method relative to the state-of-the-art methods. The code is publicly available at https://github.com/PigYang-ops/CycleGAN-QSDL. • Supervised learning methods for super-resolution PET require pairs of low- and high-resolution (LR and HR) PET images. • Quasi-supervised learning is a new type of weakly-supervised learning that can recover HR PET images from LR counterparts by leveraging similarity between unpaired LR and HR image patches. • Our proposed method can be implemented by designing a new network or modifying an existing network. • The experimental results qualitatively and quantitatively show the merits of our method relative to the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Autosomal-dominant polycystic kidney disease is a prevalent genetic disorder characterized by the development of renal cysts, leading to kidney enlargement and renal failure. Accurate measurement of total kidney volume through polycystic kidney segmentation is crucial to assess disease severity, predict progression and evaluate treatment effects. Traditional manual segmentation suffers from intra- and inter-expert variability, prompting the exploration of automated approaches. In recent years, convolutional neural networks have been employed for polycystic kidney segmentation from magnetic resonance images. However, the use of Transformer-based models, which have shown remarkable performance in a wide range of computer vision and medical image analysis tasks, remains unexplored in this area. With their self-attention mechanism, Transformers excel in capturing global context information, which is crucial for accurate organ delineations. In this paper, we evaluate and compare various convolutional-based, Transformers-based, and hybrid convolutional/Transformers-based networks for polycystic kidney segmentation. Additionally, we propose a dual-task learning scheme, where a common feature extractor is followed by per-kidney decoders, towards better generalizability and efficiency. We extensively evaluate various architectures and learning schemes on a heterogeneous magnetic resonance imaging dataset collected from 112 patients with polycystic kidney disease. Our results highlight the effectiveness of Transformer-based models for polycystic kidney segmentation and the relevancy of exploiting dual-task learning to improve segmentation accuracy and mitigate data scarcity issues. A promising ability in accurately delineating polycystic kidneys is especially shown in the presence of heterogeneous cyst distributions and adjacent cyst-containing organs. This work contribute to the advancement of reliable delineation methods in nephrology, paving the way for a broad spectrum of clinical applications. • Investigations focus on polycystic kidney delineation from magnetic resonance images. • Various convolutional and Transformer-based kidney segmentation networks are compared • Dual-task learning through a single feature extractor and per-kidney decoders is used. • Transformers show promise in delineating kidneys with diverse cyst distributions. • Dual-task learning enhances kidney segmentation accuracy in polycystic kidney disease. [ABSTRACT FROM AUTHOR]

Left atrial appendage (LAA) is the source of thrombi formation in more than 90% of strokes in patients with nonvalvular atrial fibrillation. Catheter-based LAA occlusion is being increasingly applied as a treatment strategy to prevent stroke. Anatomical complexity of LAA makes percutaneous occlusion commonly performed under transesophageal echocardiography (TEE) and X-ray (XR) guidance especially challenging. Image fusion techniques integrating 3D anatomical models derived from pre-procedural imaging into the live XR fluoroscopy can be applied to guide each step of the LAA closure. Cardiac magnetic resonance (CMR) imaging gains in importance for radiation-free evaluation of cardiac morphology as alternative to gold-standard TEE or computed tomography angiography (CTA). Manual delineation of cardiac structures from non-contrast enhanced CMR is, however, labor-intensive, tedious, and challenging due to the rather low contrast. Additionally, arrhythmia often impairs the image quality in ECG synchronized acquisitions causing blurring and motion artifacts. Thus, for cardiac segmentation in arrhythmic patients, there is a strong need for an automated image segmentation method. Deep learning-based methods have shown great promise in medical image analysis achieving superior performance in various imaging modalities and different clinical applications. Fully-convolutional neural networks (CNNs), especially U-Net, have become the method of choice for cardiac segmentation. In this paper, we propose an approach for automatic segmentation of cardiac structures from non-contrast enhanced CMR images of arrhythmic patients based on CNNs implemented in a multi-stage pipeline. Two-stage implementation allows subdividing the task into localization of the relevant cardiac structures and segmentation of these structures from the cropped sub-regions obtained from previous step leading to efficient and effective way of automated cardiac segmentation. • Deep learning approaches utilizing CNNs successfully achieved state-of-the-art for cardiac image segmentation. • Proposed two-stage CNN-based approach subdivides the segmentation task into localization and subsequent segmentation. • Introduced approach yields efficient method of cardiac image segmentation in arrhythmic patients with improved performance. [ABSTRACT FROM AUTHOR]

Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI) can be used as a non-invasive method for the assessment of myocardial perfusion. The acquired images can be utilised to analyse the spatial extent and severity of myocardial ischaemia (regions with impaired microvascular blood flow). In the present paper, we propose a novel generalisable spatio-temporal hierarchical Bayesian model (GST-HBM) to automate the detection of ischaemic lesions and improve the in silico prediction accuracy by systematically integrating spatio-temporal context information. We present a computational inference procedure with an adequate trade-off between accuracy and computational efficiency, whereby model parameters are sampled from the posterior distribution with Gibbs sampling, while lower-level hyperparameters are selected using model selection strategies based on the Watanabe Akaike information criterion (WAIC). We have assessed our method on both synthetic (in silico) data with known gold-standard and 12 sets of clinical first-pass myocardial perfusion DCE-MRI datasets. We have also carried out a comparative performance evaluation with four established alternative methods: Gaussian mixture model (GMM), opening and closing operations based on Gaussian mixture model (GMM C&O max), Markov random field constrained Gaussian mixture model (GMM-MRF) and model-based hierarchical Bayesian model (M-HBM). Our results show that the proposed GST-HBM method achieves much higher in silico prediction accuracy than the established alternative methods. Furthermore, this method appears to provide a more robust delineation of ischaemic lesions in datasets affected by spatially variant noise. • We detect myocardial ischaemia from DCE-MRI with a new hierarchical Bayesian model. • Improved robustness by using spatio-temporal context via Markov random fields. • Fully generalisable: it can employ a plurality of distance or temporal models. • Achieve higher classification accuracy than established alternative methods. [ABSTRACT FROM AUTHOR]

Digital whole slides images contain an enormous amount of information providing a strong motivation for the development of automated image analysis tools. Particularly deep neural networks show high potential with respect to various tasks in the field of digital pathology. However, a limitation is given by the fact that typical deep learning algorithms require (manual) annotations in addition to the large amounts of image data, to enable effective training. Multiple instance learning exhibits a powerful tool for training deep neural networks in a scenario without fully annotated data. These methods are particularly effective in the domain of digital pathology, due to the fact that labels for whole slide images are often captured routinely, whereas labels for patches, regions, or pixels are not. This potential resulted in a considerable number of publications, with the vast majority published in the last four years. Besides the availability of digitized data and a high motivation from the medical perspective, the availability of powerful graphics processing units exhibits an accelerator in this field. In this paper, we provide an overview of widely and effectively used concepts of (deep) multiple instance learning approaches and recent advancements. We also critically discuss remaining challenges as well as future potential. • A review of publications on multiple instance learning in digital pathology. • Due to a strong increase since 2019, we strongly focus on the last three years. • We structured according to the technically innovative building blocks. • We provide a discussion including future potential and limitations. [ABSTRACT FROM AUTHOR]

Modern cancer diagnostics involves extracting tissue specimens from suspicious areas and conducting histotechnical procedures to prepare a digitized glass slide, called Whole Slide Image (WSI), for further examination. These procedures frequently introduce different types of artifacts in the obtained WSI, and histological artifacts might influence Computational Pathology (CPATH) systems further down to a diagnostic pipeline if not excluded or handled. Deep Convolutional Neural Networks (DCNNs) have achieved promising results for the detection of some WSI artifacts, however, they do not incorporate uncertainty in their predictions. This paper proposes an uncertainty-aware Deep Kernel Learning (DKL) model to detect blurry areas and folded tissues, two types of artifacts that can appear in WSIs. The proposed probabilistic model combines a CNN feature extractor and a sparse Gaussian Processes (GPs) classifier, which improves the performance of current state-of-the-art artifact detection DCNNs and provides uncertainty estimates. We achieved 0.996 and 0.938 F1 scores for blur and folded tissue detection on unseen data, respectively. In extensive experiments, we validated the DKL model on unseen data from external independent cohorts with different staining and tissue types, where it outperformed DCNNs. Interestingly, the DKL model is more confident in the correct predictions and less in the wrong ones. The proposed DKL model can be integrated into the preprocessing pipeline of CPATH systems to provide reliable predictions and possibly serve as a quality control tool. • The presence of artifacts may hamper the performance of computational pathology systems and result in unreliable predictions. • We detect blur and folded tissue artifacts in histopathological images using our proposed Deep Kernel Learning (DKL) models. • Our DKL models are robust and demonstrate higher classification performance than the existing state-of-the-art DCNNs, and generalize well to unseen and external data. • Our DKL models provide uncertainty estimates in their predictions, which shows our DKL models are more confident in correct predictions and less in wrong ones. [Display omitted] [ABSTRACT FROM AUTHOR]

Regularization-based methods are commonly used for image registration. However, fixed regularizers have limitations in capturing details and describing the dynamic registration process. To address this issue, we propose a time multiscale registration framework for nonlinear image registration in this paper. Our approach replaces the fixed regularizer with a monotone decreasing sequence, and iteratively uses the residual of the previous step as the input for registration. Particularly, first, we introduce a dynamically varying regularization strategy that updates regularizers at each iteration and incorporates them with a multiscale framework. This approach guarantees an overall smooth deformation field in the initial stage of registration and fine-tunes local details as the images become more similar. We then deduce convergence analysis under certain conditions on the regularizers and parameters. Further, we introduce a TV-like regularizer to demonstrate the efficiency of our method. Finally, we compare our proposed multiscale algorithm with some existing methods on both synthetic images and pulmonary computed tomography (CT) images. The experimental results validate that our proposed algorithm outperforms the compared methods, especially in preserving details during image registration with sharp structures. • Incorporating dynamically varying regularizers to describe the registration process. • Extending fixed-scale deformation to multiscale to capture finer-scale deformation. • Constructing TV-like regularizers handle both smooth and discontinuous deformations. • Providing convergence analysis of the proposed time multiscale approach. [ABSTRACT FROM AUTHOR]

Micro-ultrasound (micro-US) is a novel 29-MHz ultrasound technique that provides 3-4 times higher resolution than traditional ultrasound, potentially enabling low-cost, accurate diagnosis of prostate cancer. Accurate prostate segmentation is crucial for prostate volume measurement, cancer diagnosis, prostate biopsy, and treatment planning. However, prostate segmentation on micro-US is challenging due to artifacts and indistinct borders between the prostate, bladder, and urethra in the midline. This paper presents MicroSegNet, a multi-scale annotation-guided transformer UNet model designed specifically to tackle these challenges. During the training process, MicroSegNet focuses more on regions that are hard to segment (hard regions), characterized by discrepancies between expert and non-expert annotations. We achieve this by proposing an annotation-guided binary cross entropy (AG-BCE) loss that assigns a larger weight to prediction errors in hard regions and a lower weight to prediction errors in easy regions. The AG-BCE loss was seamlessly integrated into the training process through the utilization of multi-scale deep supervision, enabling MicroSegNet to capture global contextual dependencies and local information at various scales. We trained our model using micro-US images from 55 patients, followed by evaluation on 20 patients. Our MicroSegNet model achieved a Dice coefficient of 0.939 and a Hausdorff distance of 2.02 mm, outperforming several state-of-the-art segmentation methods, as well as three human annotators with different experience levels. Our code is publicly available at https://github.com/mirthAI/MicroSegNet and our dataset is publicly available at https://zenodo.org/records/10475293. • First deep learning model for automated prostate segmentation on micro-ultrasound. • A novel annotation-guided segmentation loss that prioritizes hard-to-segment regions. • A dataset with micro-ultrasound and human prostate annotations is publicly available. [ABSTRACT FROM AUTHOR]

Automated semantic segmentation of histopathological images is an essential task in Computational Pathology (CPATH). The main limitation of Deep Learning (DL) to address this task is the scarcity of expert annotations. Crowdsourcing (CR) has emerged as a promising solution to reduce the individual (expert) annotation cost by distributing the labeling effort among a group of (non-expert) annotators. Extracting knowledge in this scenario is challenging, as it involves noisy annotations. Jointly learning the underlying (expert) segmentation and the annotators' expertise is currently a commonly used approach. Unfortunately, this approach is frequently carried out by learning a different neural network for each annotator, which scales poorly when the number of annotators grows. For this reason, this strategy cannot be easily applied to real-world CPATH segmentation. This paper proposes a new family of methods for CR segmentation of histopathological images. Our approach consists of two coupled networks: a segmentation network (for learning the expert segmentation) and an annotator network (for learning the annotators' expertise). We propose to estimate the annotators' behavior with only one network that receives the annotator ID as input, achieving scalability on the number of annotators. Our family is composed of three different models for the annotator network. Within this family, we propose a novel modeling of the annotator network in the CR segmentation literature, which considers the global features of the image. We validate our methods on a real-world dataset of Triple Negative Breast Cancer images labeled by several medical students. Our new CR modeling achieves a Dice coefficient of 0.7827, outperforming the well-known STAPLE (0.7039) and being competitive with the supervised method with expert labels (0.7723). The code is available at https://github.com/wizmik12/CRowd_Seg. • We introduce a novel segmentation model to learn from crowdsourced labels. • First model that jointly learns the segmentation and the annotators' biases in CPATH. • Our model outperforms SOTA methods that aggregate the crowdsourcing labels. • Crowdsourcing segmentation is competitive with the one trained with expert labels. • We validate the results by reporting DICE metrics and with expert pathologists. [ABSTRACT FROM AUTHOR]

Accurate and reliable segmentation of Gross Target Volume (GTV) is critical in cancer Radiation Therapy (RT) planning, but manual delineation is time-consuming and subject to inter-observer variations. Recently, deep learning methods have achieved remarkable success in medical image segmentation. However, due to the low image contrast and extreme pixel imbalance between GTV and adjacent tissues, most existing methods usually obtained limited performance on automatic GTV segmentation. In this paper, we propose a Heterogeneous Cascade Framework (HCF) from a decoupling perspective, which decomposes the GTV segmentation into independent recognition and segmentation subtasks. The former aims to screen out the abnormal slices containing GTV, while the latter performs pixel-wise segmentation of these slices. With the decoupled two-stage framework, we can efficiently filter normal slices to reduce false positives. To further improve the segmentation performance, we design a multi-level Spatial Alignment Network (SANet) based on the feature pyramid structure, which introduces a spatial alignment module into the decoder to compensate for the information loss caused by downsampling. Moreover, we propose a Combined Regularization (CR) loss and Balance-Sampling Strategy (BSS) to alleviate the pixel imbalance problem and improve network convergence. Extensive experiments on two public datasets of StructSeg2019 challenge demonstrate that our method outperforms state-of-the-art methods, especially with significant advantages in reducing false positives and accurately segmenting small objects. The code is available at https://github.com/shijun18/GTV_AutoSeg. • The decoupled two-stage segmentation framework can efficiently reduce false positives. • Spatial alignment module can compensate for the information loss caused by downsampling. • A novel combined regularization loss is proposed to alleviate the pixel imbalance problem. [ABSTRACT FROM AUTHOR]

• A new convolutional neural network is presented that takes as input an axial slice from a CT image at L3 or T4 level and generates the muscle segmentation mask of the image in almost real time (it takes less than one second (∼200 ms) for the trained network to generate the muscle mask). • The performance of the network on three large datasets is evaluated and demonstrates high Jaccard scores in the range of 0.96–0.98 on these datasets. • We validated the model's robustness by reporting the performance of the model on three different (and unseen) datasets generated by different CT devices and data acquisition settings, males and females, with a variety of muscle tissue shape and form, in various cancers attesting to the generalizability of the result. • In total more than 9000 L3 images were used for investigating (train/test) the proposed model. This is considerably higher than the number of images used for validating the methods in other papers in the literature attesting to robustness of the results. • The model trained on the large set of L3 is fine-tuned for T4 muscle segmentation. The model's performance is investigated with various experiments considering different ratio of test and training images and we find that even with a small number of images at the T4 level, having a model trained at the L3 level provides a very strong initialization to develop an accurate model for the T4 level. This indicates future generalizability to other axial locations in the CT image stack. In diseases such as cancer, patients suffer from degenerative loss of skeletal muscle (cachexia). Muscle wasting and loss of muscle function/performance (sarcopenia) can also occur during advanced aging. Assessing skeletal muscle mass in sarcopenia and cachexia is therefore of clinical interest for risk stratification. In comparison with fat, body fluids and bone, quantifying the skeletal muscle mass is more challenging. Computed tomography (CT) is one of the gold standard techniques for cancer diagnostics and analysis of progression, and therefore a valuable source of imaging for in vivo quantification of skeletal muscle mass. In this paper, we design a novel deep neural network-based algorithm for the automated segmentation of skeletal muscle in axial CT images at the third lumbar (L3) and the fourth thoracic (T4) levels. A two-branch network with two training steps is investigated. The network's performance is evaluated for three trained models on separate datasets. These datasets were generated by different CT devices and data acquisition settings. To ensure the model's robustness, each trained model was tested on all three available test sets. Errors and the effect of labeling protocol in these cases were analyzed and reported. The best performance of the proposed algorithm was achieved on 1327 L3 test samples with an overlap Jaccard score of 98% and sensitivity and specificity greater than 99%. [ABSTRACT FROM AUTHOR]

• This paper proposes a 2D/3D approach for pancreas segmentation in multimodality radiological scans (image volumes). • A novel post-processing stage, comprising of three levels, improves final tissue classification through analysis of distinct contours, and every increasing level progressively targets surrounding tissue that is located in closer proximity to the pancreas. • The proposed approach produces detailed boundary preservation and greater consistency in spatial smoothness, and tissue classification among successive slices in the image volume. • Segmentation accuracy results show robust statistical stability across multiple modalities (i.e., CT and MRI) and across datasets that were obtained using different scanner imaging protocols. • The proposed approach can be applied to other abdominal MRI and CT sequences and also, generalisable to other organ or muscular tissue segmentation tasks. Automatic pancreas segmentation in 3D radiological scans is a critical, yet challenging task. As a prerequisite for computer-aided diagnosis (CADx) systems, accurate pancreas segmentation could generate both quantitative and qualitative information towards establishing the severity of a condition, and thus provide additional guidance for therapy planning. Since the pancreas is an organ of high inter-patient anatomical variability, previous segmentation approaches report lower quantitative accuracy scores in comparison to abdominal organs such as the liver or kidneys. This paper presents a novel approach for automatic pancreas segmentation in magnetic resonance imaging (MRI) and computer tomography (CT) scans. This method exploits 3D segmentation that, when coupled with geometrical and morphological characteristics of abdominal tissue, classifies distinct contours in tight pixel-range proximity as "pancreas" or "non-pancreas". There are three main stages to this approach: (1) identify a major pancreas region and apply contrast enhancement to differentiate between pancreatic and surrounding tissue; (2) perform 3D segmentation via continuous max-flow and min-cuts approach, structured forest edge detection, and a training dataset of annotated pancreata; (3) eliminate non-pancreatic contours from resultant segmentation via morphological operations on area, structure and connectivity between distinct contours. The proposed method is evaluated on a dataset containing 82 CT image volumes, achieving mean Dice Similarity coefficient (DSC) of 79.3 ± 4.4%. Two MRI datasets containing 216 and 132 image volumes are evaluated, achieving mean DSC 79.6 ± 5.7% and 81.6 ± 5.1% respectively. This approach is statistically stable, reflected by lower metrics in standard deviation in comparison to state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

• Plus disease is a major indicator of treatment-grade retinopathy of prematurity. • The segmentation stage efficiently extracts blood vessels of varying width and length. • Using the features leaf node count and vessel density improves plus disease detection. • The proposed system would be highly useful in a tele-ophthalmology environment. Retinopathy of Prematurity (ROP) is a blinding disease affecting the retina of low birth-weight preterm infants. Accurate diagnosis of ROP is essential to identify treatment-requiring ROP, which would help to prevent childhood blindness. Plus disease, which characterizes abnormal twisting, widening and branching of the blood vessels, is a significant symptom of treatment requiring ROP. In this paper, we have developed and evaluated a computer-based analysis system for objective assessment of plus disease in ROP, which best mimics the clinical method of disease diagnosis by identifying unique vessel based features. The proposed system consists of an initial segmentation stage, which will efficiently extract blood vessels of varying width and length by utilizing structure adaptive filtering, connectivity analysis and image fusion. The paper proposes the usage of additional retinal features namely leaf node count and vessel density, to portray the abnormal growth and branching of the blood vessels and to complement the commonly used features namely tortuosity and width. The test results show a better classification of plus disease in terms of sensitivity (95%) and specificity (93%), emphasizing the superiority of the proposed segmentation algorithm and vessel-based features. An additional advantage of the proposed system is that the process of selection of relevant vessels for feature extraction is fully automated, which makes the system highly useful to the non-physician graders, owing to the unavailability of a sufficient number of ROP specialists. [ABSTRACT FROM AUTHOR]

Highlights • A methodology capable of segmenting both free-lying and clumps of abnormal cells with high overlapping rate from digitized images of conventional Pap smears. • The segmentation algorithm applies a convolutional neural network trained with more than a million patch-images to identify abnormal cell regions. • We used the average area of segmented regions to rank images according to the probability that they contain abnormal cells. • Results showed that the proposed methodology achieved more accurate results and performed faster when compared with other existing methods. • All the codes and the database will be published upon paper acceptance. Abstract Ninety years after its invention, the Pap test continues to be the most used method for the early identification of cervical precancerous lesions. In this test, the cytopathologists look for microscopic abnormalities in and around the cells, which is a time-consuming and prone to human error task. This paper introduces computational tools for cytological analysis that incorporate cell segmentation deep learning techniques. These techniques are capable of processing both free-lying and clumps of abnormal cells with a high overlapping rate from digitized images of conventional Pap smears. Our methodology employs a preprocessing step that discards images with a low probability of containing abnormal cells without prior segmentation and, therefore, performs faster when compared with the existing methods. Also, it ranks outputs based on the likelihood of the images to contain abnormal cells. We evaluate our methodology on an image database of conventional Pap smears from real scenarios, with 108 fields-of-view containing at least one abnormal cell and 86 containing only normal cells, corresponding to millions of cells. Our results show that the proposed approach achieves accurate results (MAP = 0.936), runs faster than existing methods, and it is robust to the presence of white blood cells, and other contaminants. [ABSTRACT FROM AUTHOR]

Highlights • This paper presents a multi-channel biomedical data and information computing and visualization system that can efficiently process large histological image acquired from high-resolution microscopes. • Our system can dynamically display a volume of interest and extract tissue information using a new layer-based data navigation and exploration scheme. • During data exploring, the actual resolution of the data loaded can be dynamically determined and updated. The large data rendering is synchronized in four display windows at each data layer, where 2D textures are extracted from the Medical volume and mapped onto the displayed clipping planes with arbitrary orientations. • To test the efficiency and scalability of this system, we performed extensive evaluations using different hardware systems and large histological color datasets and experimental results demonstrated a solid performance. • Our system can deliver interactive data navigation speed and display detailed imaging information in real time, which is beyond the capability of commonly available biomedical data exploration software platforms. • Taking advantage of both CPU (central processing unit) main memory and GPU (graphics processing unit) graphics memory, this software platform can efficiently compute, process and visualize very large biomedical data and enhance data information. Abstract Background and objective Modern microscopes can acquire multi-channel large histological data from tissues of human beings or animals, which contain rich biomedical information for disease diagnosis and biological feature analysis. However, due to the large size, fuzzy tissue structure, and complicated multiple elements integrated in the image color space, it is still a challenge for current software systems to effectively calculate histological data, show the inner tissue structures and unveil hidden biomedical information. Therefore, we developed new algorithms and a software platform to address this issue. Methods This paper presents a multi-channel biomedical data computing and visualization system that can efficiently process large 3D histological images acquired from high-resolution microscopes. A novelty of our system is that it can dynamically display a volume of interest and extract tissue information using a layer-based data navigation scheme. During the data exploring process, the actual resolution of the loaded data can be dynamically determined and updated, and data rendering is synchronized in four display windows at each data layer, where 2D textures are extracted from the imaging volume and mapped onto the displayed clipping planes in 3D space. Results To test the efficiency and scalability of this system, we performed extensive evaluations using several different hardware systems and large histological color datasets acquired from a CryoViz 3D digital system. The experimental results demonstrated that our system can deliver interactive data navigation speed and display detailed imaging information in real time, which is beyond the capability of commonly available biomedical data exploration software platforms. Conclusion Taking advantage of both CPU (central processing unit) main memory and GPU (graphics processing unit) graphics memory, the presented software platform can efficiently compute, process and visualize very large biomedical data and enhance data information. The performance of this system can satisfactorily address the challenges of navigating and interrogating volumetric multi-spectral large histological image at multiple resolution levels. [ABSTRACT FROM AUTHOR]

Cylindrical organs, e.g., blood vessels, airways, and intestines, are ubiquitous structures in biomedical optical imaging analysis. Image segmentation of these structures serves as a vital step in tissue physiology analysis. Traditional model-driven segmentation methods seek to fit the structure by constructing a corresponding topological geometry based on domain knowledge. Classification-based deep learning methods neglect the geometric features of the cylindrical structure and therefore cannot ensure the continuity of the segmentation surface. In this paper, by treating the cylindrical structures as a 3D graph, we introduce a novel contour-based graph neural network for 3D cylindrical structure segmentation in biomedical optical imaging. Our proposed method, which we named CylinGCN, adopts a novel learnable framework that extracts semantic features and complex topological relationships in the 3D volumetric data to achieve continuous and effective 3D segmentation. Our CylinGCN consists of a multiscale 3D semantic feature extractor for extracting inter-frame multiscale semantic features, and a residual graph convolutional network (GCN) contour generator that combines the semantic features and cylindrical topological priors to generate segmentation contours. We tested the CylinGCN framework on two types of optical tomographic imaging data, small animal whole body photoacoustic tomography (PAT) and endoscopic airway optical coherence tomography (OCT), and the results show that CylinGCN achieves state-of-the-art performance. Code will be released at https://github.com/lzc-smu/CylinGCN.git. • First contour-based model for cylinder-object segmentation in bio-optical imaging. • CylinGCN fuses local semantic and object topology for accurate segmentation. • Attention mechanism, customized graph structure and loss functions are introduced. • CylinGCN achieves state-of-the-art performance in PAT and OCT imaging datasets. [ABSTRACT FROM AUTHOR]

Accurate lymph node size estimation is critical for staging cancer patients, initial therapeutic management, and assessing response to therapy. Current standard practice for quantifying lymph node size is based on a variety of criteria that use uni-directional or bi-directional measurements. Segmentation in 3D can provide more accurate evaluations of the lymph node size. Fully convolutional neural networks (FCNs) have achieved state-of-the-art results in segmentation for numerous medical imaging applications, including lymph node segmentation. Adoption of deep learning segmentation models in clinical trials often faces numerous challenges. These include lack of pixel-level ground truth annotations for training, generalizability of the models on unseen test domains due to the heterogeneity of test cases and variation of imaging parameters. In this paper, we studied and evaluated the performance of lymph node segmentation models on a dataset that was completely independent of the one used to create the models. We analyzed the generalizability of the models in the face of a heterogeneous dataset and assessed the potential effects of different disease conditions and imaging parameters. Furthermore, we systematically compared fully-supervised and weakly-supervised methods in this context. We evaluated the proposed methods using an independent dataset comprising 806 mediastinal lymph nodes from 540 unique patients. The results show that performance achieved on the independent test set is comparable to that on the training set. Furthermore, neither the underlying disease nor the heterogeneous imaging parameters impacted the performance of the models. Finally, the results indicate that our weakly-supervised method attains 90%− 91% of the performance achieved by the fully supervised training. • Proposed lymph node segmenters generalized well on an unseen heterogeneous dataset. • Patient disease and imaging parameters did not impact the segmentation performance. • Weakly supervision achieved comparable performance to fully supervised training. [ABSTRACT FROM AUTHOR]

Medical images sometimes contain diseased regions that are different sizes and. shapes, which makes it difficult to accurately segment these areas or their edges. However, directly coupling CNN and MLP to construct global and local dependency. models may also cause significant computational complexity issues. In this paper, a. unique, lightweight UNeXt network segmentation model for medical images based on. dynamic aggregation tokens was proposed. Firstly, the Wave Block module in Wave-MLP was introduced to replace the Tok-MLP module in UNeXt. The phase term in Wave Block can dynamically aggregate tokens, improving the segmentation accuracy of the model. Secondly, an AG attention gate module is added at the skip connection to suppress irrelevant feature representations in the sampling path of the encoding. network, thereby reducing computational costs and paying attention to noise and artifacts. Finally, the Focal Tversky Loss was added to handle both binary and multiple classification jobs. Quantitative and qualitative experiments were conducted on two public datasets: COVID-19 CT and BraTS 2018 MRI. The Dice score, Precision score, recall score, and Iou score of the proposed model on the COVID-19 dataset were 0.928, 0.867, 0.916, and 0.940, respectively. On BraTS 2018, the Dice scores of the ET, WT, and TC categories were 0.933, 0.925, and 0.918, respectively, and the HD scores were 1.595, 2.348, and 1.549, respectively. At the same time, the model is lightweight and has a considerably decreased training time with GFLOPs and Params of 0.52 and 0.76, respectively. The proposed lightweight model is superior to other existing methods in terms of segmentation accuracy and computing complexity according to experimental data. • The introduction of the Wave Block module in Wave MLP replaced the tokenized MLP (Tok MLP) module in UNeXt. • The AG module attention gate was introduced at the skip connection to suppress irrelevant feature representations. • A new Loss function Focal Tversky Loss (FTL) was added, to solve the problem of class imbalance in medical image datasets. [ABSTRACT FROM AUTHOR]

Low-dose and fast PET imaging (low-count PET) play a significant role in enhancing patient safety, healthcare efficiency, and patient comfort during medical imaging procedures. To achieve high-quality images with low-count PET scans, effective reconstruction models are crucial for denoising and enhancing image quality. The main goal of this paper is to develop an effective and accurate deep learning-based method for reconstructing low-count PET images, which is a challenging problem due to the limited amount of available data and the high level of noise in the acquired images. The proposed method aims to improve the quality of reconstructed PET images while preserving important features, such as edges and small details, by combining the strengths of UNET and Transformer networks. The proposed TrUNET-MAPEM model integrates a residual UNET-transformer regularizer into the unrolled maximum a posteriori expectation maximization (MAPEM) algorithm for PET image reconstruction. A loss function based on a combination of structural similarity index (SSIM) and mean squared error (MSE) is utilized to evaluate the accuracy of the reconstructed images. The simulated dataset was generated using the Brainweb phantom, while the real patient dataset was acquired using a Siemens Biograph mMR PET scanner. We also implemented state-of-the-art methods for comparison purposes: OSEM, MAPOSEM, and supervised learning using 3D-UNET network. The reconstructed images are compared to ground truth images using metrics such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and relative root mean square error (rRMSE) to quantitatively evaluate the accuracy of the reconstructed images. Our proposed TrUNET-MAPEM approach was evaluated using both simulated and real patient data. For the patient data, our model achieved an average PSNR of 33.72 dB, an average SSIM of 0.955, and an average rRMSE of 0.39. These results outperformed other methods which had average PSNRs of 36.89 dB, 34.12 dB, and 33.52 db, average SSIMs of 0.944, 0.947, and 0.951, and average rRMSEs of 0.59, 0.49, and 0.42. For the simulated data, our model achieved an average PSNR of 31.23 dB, an average SSIM of 0.95, and an average rRMSE of 0.55. These results also outperformed other state-of-the-art methods, such as OSEM, MAPOSEM, and 3DUNET-MAPEM. The model demonstrates the potential for clinical use by successfully reconstructing smooth images while preserving edges. The comparison with other methods demonstrates the superiority of our approach, as it outperforms all other methods for all three metrics. The proposed TrUNET-MAPEM model presents a significant advancement in the field of low-count PET image reconstruction. The results demonstrate the potential for clinical use, as the model can produce images with reduced noise levels and better edge preservation compared to other reconstruction and post-processing algorithms. The proposed approach may have important clinical applications in the early detection and diagnosis of various diseases. • Effective and accurate 3D deep learning-based method for reconstructing low-count PET images. • Enhance PET image reconstruction quality while preserving edges and details, by combining UNET and Transformer networks. • High accuracy in terms of PSNR, SSIM, and rRMSE on both simulated and clinical datasets. [ABSTRACT FROM AUTHOR]

In this paper, we address the problem of estimating remaining surgery duration (RSD) from surgical video frames. We propose a Bayesian long short-term memory (LSTM) network-based Deep Negative Correlation Learning approach called BD-Net for accurate regression of RSD prediction as well as estimation of prediction uncertainty. Our method aims to extract discriminative visual features from surgical video frames and model the temporal dependencies among frames to improve the RSD prediction accuracy. To this end, we propose to train an ensemble of Bayesian LSTMs on top of a backbone network by the way of deep negative correlation learning (DNCL). More specifically, we deeply learn a pool of decorrelated Bayesian regressors with sound generalization capabilities through managing their intrinsic diversities. BD-Net is simple and efficient. After training, it can produce both RSD prediction and uncertainty estimation in a single inference run. We demonstrate the efficacy of BD-Net on publicly available datasets of two different types of surgeries: one containing 101 cataract microscopic surgeries with short durations and the other containing 80 cholecystectomy laparoscopic surgeries with relatively longer durations. Experimental results on both datasets demonstrate that the proposed BD-Net achieves better results than the state-of-the-art (SOTA) methods. A reference implementation of our method can be found at: https://github.com/jywu511/BD-Net. • A novel Bayesian LSTM-based deep negative correlation learning network called BD-NET; • BD-Net can estimate both remaining surgery duration and prediction uncertainty; • Comprehensive validation on two publicly available datasets. [ABSTRACT FROM AUTHOR]

Multimodal images such as magnetic resonance imaging (MRI) and positron emission tomography (PET) could provide complementary information about the brain and have been widely investigated for the diagnosis of neurodegenerative disorders such as Alzheimer's disease (AD). However, multimodal brain images are often incomplete in clinical practice. It is still challenging to make use of multimodality for disease diagnosis with missing data. In this paper, we propose a deep learning framework with the multi-level guided generative adversarial network (MLG-GAN) and multimodal transformer (Mul-T) for incomplete image generation and disease classification, respectively. First, MLG-GAN is proposed to generate the missing data, guided by multi-level information from voxels, features, and tasks. In addition to voxel-level supervision and task-level constraint, a feature-level auto-regression branch is proposed to embed the features of target images for an accurate generation. With the complete multimodal images, we propose a Mul-T network for disease diagnosis, which can not only combine the global and local features but also model the latent interactions and correlations from one modality to another with the cross-modal attention mechanism. Comprehensive experiments on three independent datasets (i.e., ADNI-1, ADNI-2, and OASIS-3) show that the proposed method achieves superior performance in the tasks of image generation and disease diagnosis compared to state-of-the-art methods. • A novel deep learning framework is proposed for incomplete image generation and multimodal diagnosis in Alzheimer's disease. • A multi-level guided generative adversarial network is proposed for generation of missing imaging data. • A multimodal transformer is proposed to model multimodal interactions and combine the features for disease classification. • Experiments on three independent datasets demonstrate the superior performance of our method [ABSTRACT FROM AUTHOR]

The performance and quality of medical procedures and treatments are inextricably linked to technological development. The application of more advanced techniques provides the opportunity to gain wider knowledge and deeper understanding of the human body and mind functioning. The eye tracking methods used to register eye movement to find the direction and targets of a person's gaze are well in line with the nature of the topic. By providing methods for capturing and processing images of the eye it has become possible not only to reveal abnormalities in eye functioning but also to conduct cognitive studies focused on learning about peoples’ emotions and intentions. The usefulness of the application of eye tracking technology in medicine was proved in many research studies. The aim of this paper is to give an insight into those studies and the way they utilize eye imaging in medical applications. These studies were differentiated taking their purpose and experimental paradigms into account. Additionally, methods for eye movement visualization and metrics for its quantifying were presented. Apart from presenting the state of the art, the aim of the paper was also to point out possible applications of eye tracking in medicine that have not been exhaustively investigated yet, and are going to be a perspective long-term direction of research. [ABSTRACT FROM AUTHOR]