Browse
Submit Data
Databases
API
Help

Dataset Information

13 Views

0 Connections

0 Citations

0 Reanalyses

0 Downloads

Omics score: 0

Attention gated networks: Learning to leverage salient regions in medical images.

ABSTRACT: We propose a novel attention gate (AG) model for medical image analysis that automatically learns to focus on target structures of varying shapes and sizes. Models trained with AGs implicitly learn to suppress irrelevant regions in an input image while highlighting salient features useful for a specific task. This enables us to eliminate the necessity of using explicit external tissue/organ localisation modules when using convolutional neural networks (CNNs). AGs can be easily integrated into standard CNN models such as VGG or U-Net architectures with minimal computational overhead while increasing the model sensitivity and prediction accuracy. The proposed AG models are evaluated on a variety of tasks, including medical image classification and segmentation. For classification, we demonstrate the use case of AGs in scan plane detection for fetal ultrasound screening. We show that the proposed attention mechanism can provide efficient object localisation while improving the overall prediction performance by reducing false positives. For segmentation, the proposed architecture is evaluated on two large 3D CT abdominal datasets with manual annotations for multiple organs. Experimental results show that AG models consistently improve the prediction performance of the base architectures across different datasets and training sizes while preserving computational efficiency. Moreover, AGs guide the model activations to be focused around salient regions, which provides better insights into how model predictions are made. The source code for the proposed AG models is publicly available.

SUBMITTER: Schlemper J

PROVIDER: S-EPMC7610718 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Json Xml

Similar Datasets

Learning Medical Materials From Radiography Images.

Project description:Deep learning models have been shown to be effective for material analysis, a subfield of computer vision, on natural images. In medicine, deep learning systems have been shown to more accurately analyze radiography images than algorithmic approaches and even experts. However, one major roadblock to applying deep learning-based material analysis on radiography images is a lack of material annotations accompanying image sets. To solve this, we first introduce an automated procedure to augment annotated radiography images into a set of material samples. Next, using a novel Siamese neural network that compares material sample pairs, called D-CNN, we demonstrate how to learn a perceptual distance metric between material categories. This system replicates the actions of human annotators by discovering attributes that encode traits that distinguish materials in radiography images. Finally, we update and apply MAC-CNN, a material recognition neural network, to demonstrate this system on a dataset of knee X-rays and brain MRIs with tumors. Experiments show that this system has strong predictive power on these radiography images, achieving 92.8% accuracy at predicting the material present in a local region of an image. Our system also draws interesting parallels between human perception of natural materials and materials in radiography images.

| S-EPMC8320745 | biostudies-literature

Natural Images Allow Universal Adversarial Attacks on Medical Image Classification Using Deep Neural Networks with Transfer Learning.

Project description:Transfer learning from natural images is used in deep neural networks (DNNs) for medical image classification to achieve a computer-aided clinical diagnosis. Although the adversarial vulnerability of DNNs hinders practical applications owing to the high stakes of diagnosis, adversarial attacks are expected to be limited because training datasets (medical images), which are often required for adversarial attacks, are generally unavailable in terms of security and privacy preservation. Nevertheless, in this study, we demonstrated that adversarial attacks are also possible using natural images for medical DNN models with transfer learning, even if such medical images are unavailable; in particular, we showed that universal adversarial perturbations (UAPs) can also be generated from natural images. UAPs from natural images are useful for both non-targeted and targeted attacks. The performance of UAPs from natural images was significantly higher than that of random controls. The use of transfer learning causes a security hole, which decreases the reliability and safety of computer-based disease diagnosis. Model training from random initialization reduced the performance of UAPs from natural images; however, it did not completely avoid vulnerability to UAPs. The vulnerability of UAPs to natural images is expected to become a significant security threat.

| S-EPMC8875959 | biostudies-literature

Deep Learning in Medical Hyperspectral Images: A Review.

Project description:With the continuous progress of development, deep learning has made good progress in the analysis and recognition of images, which has also triggered some researchers to explore the area of combining deep learning with hyperspectral medical images and achieve some progress. This paper introduces the principles and techniques of hyperspectral imaging systems, summarizes the common medical hyperspectral imaging systems, and summarizes the progress of some emerging spectral imaging systems through analyzing the literature. In particular, this article introduces the more frequently used medical hyperspectral images and the pre-processing techniques of the spectra, and in other sections, it discusses the main developments of medical hyperspectral combined with deep learning for disease diagnosis. On the basis of the previous review, tne limited factors in the study on the application of deep learning to hyperspectral medical images are outlined, promising research directions are summarized, and the future research prospects are provided for subsequent scholars.

| S-EPMC9784550 | biostudies-literature

Identifying Periampullary Regions in MRI Images Using Deep Learning.

Project description:BackgroundDevelopment and validation of a deep learning method to automatically segment the peri-ampullary (PA) region in magnetic resonance imaging (MRI) images.MethodsA group of patients with or without periampullary carcinoma (PAC) was included. The PA regions were manually annotated in MRI images by experts. Patients were randomly divided into one training set, one validation set, and one test set. Deep learning methods were developed to automatically segment the PA region in MRI images. The segmentation performance of the methods was compared in the validation set. The model with the highest intersection over union (IoU) was evaluated in the test set.ResultsThe deep learning algorithm achieved optimal accuracies in the segmentation of the PA regions in both T1 and T2 MRI images. The value of the IoU was 0.68, 0.68, and 0.64 for T1, T2, and combination of T1 and T2 images, respectively.ConclusionsDeep learning algorithm is promising with accuracies of concordance with manual human assessment in segmentation of the PA region in MRI images. This automated non-invasive method helps clinicians to identify and locate the PA region using preoperative MRI scanning.

| S-EPMC8193851 | biostudies-literature

Autosegmentation of Prostate Zones and Cancer Regions from Biparametric Magnetic Resonance Images by Using Deep-Learning-Based Neural Networks.

Project description:The accuracy in diagnosing prostate cancer (PCa) has increased with the development of multiparametric magnetic resonance imaging (mpMRI). Biparametric magnetic resonance imaging (bpMRI) was found to have a diagnostic accuracy comparable to mpMRI in detecting PCa. However, prostate MRI assessment relies on human experts and specialized training with considerable inter-reader variability. Deep learning may be a more robust approach for prostate MRI assessment. Here we present a method for autosegmenting the prostate zone and cancer region by using SegNet, a deep convolution neural network (DCNN) model. We used PROSTATEx dataset to train the model and combined different sequences into three channels of a single image. For each subject, all slices that contained the transition zone (TZ), peripheral zone (PZ), and PCa region were selected. The datasets were produced using different combinations of images, including T2-weighted (T2W) images, diffusion-weighted images (DWI) and apparent diffusion coefficient (ADC) images. Among these groups, the T2W + DWI + ADC images exhibited the best performance with a dice similarity coefficient of 90.45% for the TZ, 70.04% for the PZ, and 52.73% for the PCa region. Image sequence analysis with a DCNN model has the potential to assist PCa diagnosis.

| S-EPMC8070192 | biostudies-literature

Exploring the effect of context and expertise on attention: is attention shifted by information in medical images?

Project description:Radiologists make critical decisions based on searching and interpreting medical images. The probability of a lung nodule differs across anatomical regions within the chest, raising the possibility that radiologists might have a prior expectation that creates an attentional bias. The development of expertise is also thought to cause "tuning" to relevant features, allowing radiologists to become faster and more accurate at detecting potential masses within their domain of expertise. Here, we tested both radiologists and control participants with a novel attentional-cueing paradigm to investigate whether the deployment of attention was affected (1) by a context that might invoke prior knowledge for experts, (2) by a nodule localized either on the same or on opposite sides as a subsequent target, and (3) by inversion of the nodule-present chest radiographs, to assess the orientation specificity of any effects. The participants also performed a nodule detection task to verify that our presentation duration was sufficient to extract diagnostic information. We saw no evidence of priors triggered by a normal chest radiograph cue affecting attention. When the cue was an upright abnormal chest radiograph, radiologists were faster when the lateralised nodule and the subsequent target appeared at the same rather than at opposite locations, suggesting attention was captured by the nodule. The opposite pattern was present for inverted images. We saw no evidence of cueing for control participants in any condition, which suggests that radiologists are indeed more sensitive to visual features that are not perceived as salient by naïve observers.

| S-EPMC6647457 | biostudies-literature

Leverage Large-Scale Biological Networks to Decipher the Genetic Basis of Human Diseases Using Machine Learning.

Project description:A fundamental question in precision medicine is to quantitatively decode the genetic basis of complex human diseases, which will enable the development of predictive models of disease risks based on personal genome sequences. To account for the complex systems within different cellular contexts, large-scale regulatory networks are critical components to be integrated into the analysis. Based on the fast accumulation of multiomics and disease genetics data, advanced machine learning algorithms and efficient computational tools are becoming the driving force in predicting phenotypes from genotypes, identifying potential causal genetic variants, and revealing disease mechanisms. Here, we review the state-of-the-art methods for this topic and describe a computational pipeline that assembles a series of algorithms together to achieve improved disease genetics prediction through the delineation of regulatory circuitry step by step.

| S-EPMC7433890 | biostudies-literature

Skin strata delineation in reflectance confocal microscopy images using recurrent convolutional networks with attention.

Project description:Reflectance confocal microscopy (RCM) is an effective non-invasive tool for cancer diagnosis. However, acquiring and reading RCM images requires extensive training and experience, and novice clinicians exhibit high discordance in diagnostic accuracy. Quantitative tools to standardize image acquisition could reduce both required training and diagnostic variability. To perform diagnostic analysis, clinicians collect a set of RCM mosaics (RCM images concatenated in a raster fashion to extend the field view) at 4-5 specific layers in skin, all localized in the junction between the epidermal and dermal layers (dermal-epidermal junction, DEJ), necessitating locating that junction before mosaic acquisition. In this study, we automate DEJ localization using deep recurrent convolutional neural networks to delineate skin strata in stacks of RCM images collected at consecutive depths. Success will guide to automated and quantitative mosaic acquisition thus reducing inter operator variability and bring standardization in imaging. Testing our model against an expert labeled dataset of 504 RCM stacks, we achieved [Formula: see text] classification accuracy and nine-fold reduction in the number of anatomically impossible errors compared to the previous state-of-the-art.

| S-EPMC8206415 | biostudies-literature

Salient objects dominate the central fixation bias when orienting toward images.

Project description:Short-latency saccades are often biased toward salient objects or toward the center of images, for example, when inspecting photographs of natural scenes. Here, we measured the contribution of salient objects and central fixation bias to visual selection over time. Participants made saccades to images containing one salient object on a structured background and were instructed to either look at (i) the image center, (ii) the salient object, or (iii) at a cued position halfway in between the two. Results revealed, first, an early involuntary bias toward the image center irrespective of strategic behavior or the location of objects in the image. Second, the salient object bias was stronger than the center bias and prevailed over the latter when they directly competed for visual selection. In a second experiment, we tested whether the center bias depends on how well the image can be segregated from the monitor background. We asked participants to explore images that either did or did not contain a salient object while we manipulated the contrast between image background and monitor background to make the image borders more or less visible. The initial orienting toward the image was not affected by the image-monitor contrast, but only by the presence of objects-with a strong bias toward the center of images containing no object. Yet, a low image-monitor contrast reduced this center bias during the subsequent image exploration.

| S-EPMC8399466 | biostudies-literature

Context Matters: Graph-based Self-supervised Representation Learning for Medical Images.

Project description:Supervised learning method requires a large volume of annotated datasets. Collecting such datasets is time-consuming and expensive. Until now, very few annotated COVID-19 imaging datasets are available. Although self-supervised learning enables us to bootstrap the training by exploiting unlabeled data, the generic self-supervised methods for natural images do not sufficiently incorporate the context. For medical images, a desirable method should be sensitive enough to detect deviation from normal-appearing tissue of each anatomical region; here, anatomy is the context. We introduce a novel approach with two levels of self-supervised representation learning objectives: one on the regional anatomical level and another on the patient-level. We use graph neural networks to incorporate the relationship between different anatomical regions. The structure of the graph is informed by anatomical correspondences between each patient and an anatomical atlas. In addition, the graph representation has the advantage of handling any arbitrarily sized image in full resolution. Experiments on large-scale Computer Tomography (CT) datasets of lung images show that our approach compares favorably to baseline methods that do not account for the context. We use the learnt embedding to quantify the clinical progression of COVID-19 and show that our method generalizes well to COVID-19 patients from different hospitals. Qualitative results suggest that our model can identify clinically relevant regions in the images.

| S-EPMC8513790 | biostudies-literature

OmicsDI is part of the ELIXIR infrastructure

OmicsDI is an Elixir interoperability service. Learn more ›

Tweets

OmicsDI Databases

PRIDE
PeptideAtlas
MassIVE
JPOST Repository
Physiome Model Repository

EGA
EVA
ENA
LINCS
PAXDB
Cell Collective

MetaboLights
Metabolomics Workbench
MetabolomeExpress
GNPS
BioModels
FAIRDOMHub

ArrayExpress
dbGaP
ExpressionAtlas
GEO
NODE

Information

Databases
Help
API
Contact us
Code on GitHub
Terms of use
Submit Data