Improving the Efficacy of Deep-Learning Models for Heart Beat Detection on Heterogeneous Datasets.
Ontology highlight
ABSTRACT:
SUBMITTER: Bizzego A
PROVIDER: S-EPMC8698903 | biostudies-literature | 2021 Nov
REPOSITORIES: biostudies-literature
Publications
Improving the Efficacy of Deep-Learning Models for Heart Beat Detection on Heterogeneous Datasets.
Bizzego Andrea A Gabrieli Giulio G Neoh Michelle Jin Yee MJY Esposito Gianluca G
Bioengineering (Basel, Switzerland) 20211128 12
Deep learning (DL) has greatly contributed to bioelectric signal processing, in particular to extract physiological markers. However, the efficacy and applicability of the results proposed in the literature is often constrained to the population represented by the data used to train the models. In this study, we investigate the issues related to applying a DL model on heterogeneous datasets. In particular, by focusing on heart beat detection from electrocardiogram signals (ECG), we show that the ...[more]
Similar Datasets
Project description:Beat-by-beat arrhythmia detection in ambulatory electrocardiogram (ECG) monitoring is critical for the evaluation and prognosis of cardiac arrhythmias, however, it is a highly professional demanding and time-consuming task. Current methods for automatic beat-by-beat arrhythmia detection suffer from poor generalization ability due to the lack of large-sample and finely-annotated (labels are given to each beat) ECG data for model training. In this work, we propose a weakly supervised deep learning framework for arrhythmia detection (WSDL-AD), which permits training a fine-grained (beat-by-beat) arrhythmia detector with the use of large amounts of coarsely annotated ECG data (labels are given to each recording) to improve the generalization ability. In this framework, heartbeat classification and recording classification are integrated into a deep neural network for end-to-end training with only recording labels. Several techniques, including knowledge-based features, masked aggregation, and supervised pre-training, are proposed to improve the accuracy and stability of the heartbeat classification under weak supervision. The developed WSDL-AD model is trained for the detection of ventricular ectopic beats (VEB) and supraventricular ectopic beats (SVEB) on five large-sample and coarsely-annotated datasets and the model performance is evaluated on three independent benchmarks according to the recommendations from the Association for the Advancement of Medical Instrumentation (AAMI). The experimental results show that our method improves the F 1 score of supraventricular ectopic beats detection by 8%-290% and the F1 of ventricular ectopic beats detection by 4%-11% on the benchmarks compared with the state-of-the-art methods of supervised learning. It demonstrates that the WSDL-AD framework can leverage the abundant coarsely-labeled data to achieve a better generalization ability than previous methods while retaining fine detection granularity. Therefore, this framework has a great potential to be used in clinical and telehealth applications. The source code is available at https://github.com/sdnjly/WSDL-AD.
| S-EPMC9037749 | biostudies-literature
Project description:Amyotrophic lateral sclerosis (ALS) is a fatal neurological disease marked by motor deterioration and cognitive decline. Early diagnosis is challenging due to the complexity of sporadic ALS and the lack of a defined risk population. In this study, we developed Miniset-DenseSENet, a convolutional neural network combining DenseNet121 with a Squeeze-and-Excitation attention mechanism, using 190 autopsy brain images from the Gregory Laboratory at the University of Aberdeen. The model distinguishes controls, ALS patients with no cognitive impairment, and ALS patients with cognitive impairment (ALS-frontotemporal dementia) with 97.37% accuracy, addressing a significant challenge in overlapping neurodegenerative disorders involving TDP-43 proteinopathy. Miniset-DenseSENet outperformed other transfer learning models, achieving a sensitivity of 1 and specificity of 0.95. These findings suggest that integrating transfer learning and attention mechanisms into neuroimaging can enhance diagnostic accuracy, enabling earlier ALS detection and improving patient stratification. This model has the potential to guide clinical decisions and support personalied therapeutic strategies.
| S-EPMC11868594 | biostudies-literature
Project description:BackgroundElectrocardiogram (ECG) signal classification plays a critical role in the automatic diagnosis of heart abnormalities. While most ECG signal patterns cannot be recognized by a human interpreter, they can be detected with precision using artificial intelligence approaches, making the ECG a powerful non-invasive biomarker. However, performing rapid and accurate ECG signal classification is difficult due to the low amplitude, complexity, and non-linearity. The widely-available deep learning (DL) method we propose has presented an opportunity to substantially improve the accuracy of automated ECG classification analysis using rhythm or beat features. Unfortunately, a comprehensive and general evaluation of the specific DL architecture for ECG analysis across a wide variety of rhythm and beat features has not been previously reported. Some previous studies have been concerned with detecting ECG class abnormalities only through rhythm or beat features separately.MethodsThis study proposes a single architecture based on the DL method with one-dimensional convolutional neural network (1D-CNN) architecture, to automatically classify 24 patterns of ECG signals through both rhythm and beat. To validate the proposed model, five databases which consisted of nine-class of ECG-base rhythm and 15-class of ECG-based beat were used in this study. The proposed DL network was applied and studied with varying datasets with different frequency samplings in intra and inter-patient scheme.ResultsUsing a 10-fold cross-validation scheme, the performance results had an accuracy of 99.98%, a sensitivity of 99.90%, a specificity of 99.89%, a precision of 99.90%, and an F1-score of 99.99% for ECG rhythm classification. Additionally, for ECG beat classification, the model obtained an accuracy of 99.87%, a sensitivity of 96.97%, a specificity of 99.89%, a precision of 92.23%, and an F1-score of 94.39%. In conclusion, this study provides clinicians with an advanced methodology for detecting and discriminating heart abnormalities between different ECG rhythm and beat assessments by using one outstanding proposed DL architecture.
| S-EPMC8802771 | biostudies-literature
Project description:MicroRNAs (miRNAs) are involved in the post-transcriptional regulation of protein abundance and thus have a great impact on the resulting phenotype. It is, therefore, no wonder that they have been implicated in many diseases ranging from virus infections to cancer. This impact on the phenotype leads to a great interest in establishing the miRNAs of an organism. Experimental methods are complicated which led to the development of computational methods for pre-miRNA detection. Such methods generally employ machine learning to establish models for the discrimination between miRNAs and other sequences. Positive training data for model establishment, for the most part, stems from miRBase, the miRNA registry. The quality of the entries in miRBase has been questioned, though. This unknown quality led to the development of filtering strategies in attempts to produce high quality positive datasets which can lead to a scarcity of positive data. To analyze the quality of filtered data we developed a machine learning model and found it is well able to establish data quality based on intrinsic measures. Additionally, we analyzed which features describing pre-miRNAs could discriminate between low and high quality data. Both models are applicable to data from miRBase and can be used for establishing high quality positive data. This will facilitate the development of better miRNA detection tools which will make the prediction of miRNAs in disease states more accurate. Finally, we applied both models to all miRBase data and provide the list of high quality hairpins.
| S-EPMC6042829 | biostudies-literature
Project description:BackgroundRadiomics promises to enhance the discriminative performance for clinically significant prostate cancer (csPCa), but still lacks validation in real-life scenarios. This study investigates the classification performance and robustness of machine learning radiomics models in heterogeneous MRI datasets to characterize suspicious prostate lesions for non-invasive prediction of prostate cancer (PCa) aggressiveness compared to conventional imaging biomarkers.MethodsA total of 142 patients with clinical suspicion of PCa underwent 1.5T or 3T biparametric MRI (7 scanner types, 14 institutions) and exhibited suspicious lesions [prostate Imaging Reporting and Data System (PI-RADS) score ≥3] in peripheral or transitional zones. Whole-gland and index-lesion segmentations were performed semi-automatically. A total of 1,482 quantitative morphologic, shape, texture, and intensity-based radiomics features were extracted from T2-weighted and apparent diffusion coefficient (ADC)-images and assessed using random forest and logistic regression models. Five-fold cross-validation performance in terms of area under the ROC curve was compared to mean ADC (mADC), PI-RADS and prostate-specific antigen density (PSAD). Bias mitigation techniques targeting the high-dimensional feature space and inherent class imbalance were applied and robustness of results was systematically evaluated.ResultsTrained models showed mean area under the curves (AUCs) ranging from 0.78 to 0.83 in csPCa classification. Despite using mitigation techniques, high performance variability of results could be demonstrated. Trained models achieved on average numerically higher classification performance compared to clinical parameters PI-RADS (AUC =0.78), mADC (AUC =0.71) and PSAD (AUC =0.63).ConclusionsRadiomics models' classification performance of csPCa was numerically but not significantly higher than PI-RADS scoring. Overall, clinical applicability in heterogeneous MRI datasets is limited because of high variability of results. Performance variability, robustness and reproducibility of radiomics-based measures should be addressed more transparently in future research to enable broad clinical application.
| S-EPMC9622454 | biostudies-literature
Project description:MotivationCells contain dozens of major organelles and thousands of other structures, many of which vary extensively in their number, size, shape and spatial distribution. This complexity and variation dramatically complicates the use of both traditional and deep learning methods to build accurate models of cell organization. Most cellular organelles are distinct objects with defined boundaries that do not overlap, while the pixel resolution of most imaging methods is n sufficient to resolve these boundaries. Thus while cell organization is conceptually object-based, most current methods are pixel-based. Using extensive image collections in which particular organelles were fluorescently labeled, deep learning methods can be used to build conditional autoencoder models for particular organelles. A major advance occurred with the use of a U-net approach to make multiple models all conditional upon a common reference, unlabeled image, allowing the relationships between different organelles to be at least partially inferred.ResultsWe have developed improved Generative Adversarial Networks-based approaches for learning these models and have also developed novel criteria for evaluating how well synthetic cell images reflect the properties of real images. The first set of criteria measure how well models preserve the expected property that organelles do not overlap. We also developed a modified loss function that allows retraining of the models to minimize that overlap. The second set of criteria uses object-based modeling to compare object shape and spatial distribution between synthetic and real images. Our work provides the first demonstration that, at least for some organelles, deep learning models can capture object-level properties of cell images.Availability and implementationhttp://murphylab.cbd.cmu.edu/Software/2022_insilico.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC9710556 | biostudies-literature
Project description:Cardiac structure contouring is a time consuming and tedious manual activity used for radiotherapeutic dose toxicity planning. We developed an automatic cardiac structure segmentation pipeline for use in low-dose non-contrast planning CT based on deep learning algorithms for small datasets. Fifty CT scans were retrospectively selected and the whole heart, ventricles and atria were contoured. A two stage deep learning pipeline was trained on 41 non contrast planning CTs, tuned with 3 CT scans and validated on 6 CT scans. In the first stage, An InceptionResNetV2 network was used to identify the slices that contained cardiac structures. The second stage consisted of three deep learning models trained on the images containing cardiac structures to segment the structures. The three deep learning models predicted the segmentations/contours on axial, coronal and sagittal images and are combined to create the final prediction. The final accuracy of the pipeline was quantified on 6 volumes by calculating the Dice similarity coefficient (DC), 95% Hausdorff distance (95% HD) and volume ratios between predicted and ground truth volumes. Median DC and 95% HD of 0.96, 0.88, 0.92, 0.80 and 0.82, and 1.86, 2.98, 2.02, 6.16 and 6.46 were achieved for the whole heart, right and left ventricle, and right and left atria respectively. The median differences in volume were -4, -1, + 5, -16 and -20% for the whole heart, right and left ventricle, and right and left atria respectively. The automatic contouring pipeline achieves good results for whole heart and ventricles. Robust automatic contouring with deep learning methods seems viable for local centers with small datasets.
| S-EPMC8956542 | biostudies-literature
Project description:Biomedical relation extraction (RE) is the task of automatically identifying and characterizing relations between biomedical concepts from free text. RE is a central task in biomedical natural language processing (NLP) research and plays a critical role in many downstream applications, such as literature-based discovery and knowledge graph construction. State-of-the-art methods were used primarily to train machine learning models on individual RE datasets, such as protein-protein interaction and chemical-induced disease relation. Manual dataset annotation, however, is highly expensive and time-consuming, as it requires domain knowledge. Existing RE datasets are usually domain-specific or small, which limits the development of generalized and high-performing RE models. In this work, we present a novel framework for systematically addressing the data heterogeneity of individual datasets and combining them into a large dataset. Based on the framework and dataset, we report on BioREx, a data-centric approach for extracting relations. Our evaluation shows that BioREx achieves significantly higher performance than the benchmark system trained on the individual dataset, setting a new SOTA from 74.4% to 79.6% in F-1 measure on the recently released BioRED corpus. We further demonstrate that the combined dataset can improve performance for five different RE tasks. In addition, we show that on average BioREx compares favorably to current best-performing methods such as transfer learning and multi-task learning. Finally, we demonstrate BioREx's robustness and generalizability in two independent RE tasks not previously seen in training data: drug-drug N-ary combination and document-level gene-disease RE. The integrated dataset and optimized method have been packaged as a stand-alone tool available at https://github.com/ncbi/BioREx.
| S-EPMC10370213 | biostudies-literature
Project description:The intuition behind this data acquisition is to encourage research for addressing the problem of weeds in agriculture through computer vision applications. Data is acquired in the form of images from uniform and random crop-spacing fields. In other words, we have taken a step forward to identify weeds from fields that follow any method of sowing, which ultimately leads to the transformation of traditional agriculture into precision agriculture. Sorghum crop and its associated weeds are chosen as the research objects during this process. These acquired data are used in framing two datasets. The first dataset termed 'SorghumWeedDataset_Classification' is a crop-weed classification dataset created with 4312 data samples for addressing crop-weed classification problems. The second dataset termed 'SorghumWeedDataset_Segmentation' is a crop-weed segmentation dataset that contains 5555 manually pixel-wise annotated data segments from 252 data samples for addressing crop-weed localization, detection, and segmentation problems. All data samples are acquired in April and May 2023 from Sri Ramaswamy Memorial (SRM) Care Farm, Chengalpattu district, Tamil Nadu, India. Manually annotated data samples and data segments are verified by agronomists. The datasets are made publicly available to the research community to solve the crop-weed problems using state-of-the-art image processing, machine learning, and deep learning algorithms. To the best of our knowledge, these are the first open-access crop-weed research datasets from Indian fields for classification and segmentation to deal with weed issues in uniform and random crop-spacing fields. However, other available datasets (from Indian fields) are either non-research datasets or available on subscription/request.
| S-EPMC10789999 | biostudies-literature
Project description:BackgroundHigh-throughput and selective detection of organelles in immunofluorescence images is an important but demanding task in cell biology. The centriole organelle is critical for fundamental cellular processes, and its accurate detection is key for analysing centriole function in health and disease. Centriole detection in human tissue culture cells has been achieved typically by manual determination of organelle number per cell. However, manual cell scoring of centrioles has a low throughput and is not reproducible. Published semi-automated methods tally the centrosome surrounding centrioles and not centrioles themselves. Furthermore, such methods rely on hard-coded parameters or require a multichannel input for cross-correlation. Therefore, there is a need for developing an efficient and versatile pipeline for the automatic detection of centrioles in single channel immunofluorescence datasets.ResultsWe developed a deep-learning pipeline termed CenFind that automatically scores cells for centriole numbers in immunofluorescence images of human cells. CenFind relies on the multi-scale convolution neural network SpotNet, which allows the accurate detection of sparse and minute foci in high resolution images. We built a dataset using different experimental settings and used it to train the model and evaluate existing detection methods. The resulting average F1-score achieved by CenFind is > 90% across the test set, demonstrating the robustness of the pipeline. Moreover, using the StarDist-based nucleus detector, we link the centrioles and procentrioles detected with CenFind to the cell containing them, overall enabling automatic scoring of centriole numbers per cell.ConclusionsEfficient, accurate, channel-intrinsic and reproducible detection of centrioles is an important unmet need in the field. Existing methods are either not discriminative enough or focus on a fixed multi-channel input. To fill this methodological gap, we developed CenFind, a command line interface pipeline that automates cell scoring of centrioles, thereby enabling channel-intrinsic, accurate and reproducible detection across experimental modalities. Moreover, the modular nature of CenFind enables its integration in other pipelines. Overall, we anticipate CenFind to prove critical for accelerating discoveries in the field.
| S-EPMC10045196 | biostudies-literature