Cardiac arrhythmia detection using deep learning approach and time frequency representation of ECG signals

BMC Medical Informatics and Decision Making, Год журнала: 2023, Номер 23(1)

Опубликована: Окт. 19, 2023

Cardiac arrhythmia is a cardiovascular disorder characterized by disturbances in the heartbeat caused electrical conduction anomalies cardiac muscle. Clinically, ECG machines are utilized to diagnose and monitor noninvasively. Since signals dynamic nature depict various complex information, visual assessment analysis time consuming very difficult. Therefore, an automated system that can assist physicians easy detection of needed.The main objective this study was create deep learning model capable accurately classifying into three categories: (ARR), congestive heart failure (CHF), normal sinus rhythm (NSR). To achieve this, data from MIT-BIH BIDMC databases available on PhysioNet were preprocessed segmented before being for training. Pretrained models, ResNet 50 AlexNet, fine-tuned configured optimal classification results. The outcome measures evaluating performance F-measure, recall, precision, sensitivity, specificity, accuracy, obtained multi-class confusion matrix.The proposed showed overall accuracy 99.2%, average sensitivity specificity 99.6%, precision F- measure 99.2% test data.The work introduced robust approach arrhythmias comparison with most recent state art will reduce diagnosis error occurs investigation signals.

Язык: Английский

---

Deep Neural Network for EEG Signal-Based Subject-Independent Imaginary Mental Task Classification

Diagnostics, Год журнала: 2023, Номер 13(4), С. 640 - 640

Опубликована: Фев. 9, 2023

BACKGROUND. Mental task identification using electroencephalography (EEG) signals is required for patients with limited or no motor movements. A subject-independent mental classification framework can be applied to identify the of a subject available training statistics. Deep learning frameworks are popular among researchers analyzing both spatial and time series data, making them well-suited classifying EEG signals. METHOD. In this paper, deep neural network model proposed an imagined from signal data. Pre-computed features were obtained after raw acquired subjects spatially filtered by applying Laplacian surface. To handle high-dimensional principal component analysis (PCA) was performed which helps in extraction most discriminating input vectors. RESULT. The non-invasive aims extract task-specific data particular subject. on average combined Power Spectrum Density (PSD) values all but one performance based (DNN) evaluated benchmark dataset. We achieved 77.62% accuracy. CONCLUSION. comparison related existing works validated that cross-subject outperforms state-of-the-art algorithm terms performing accurate

Язык: Английский

---

Deep Convolutional Neural Networks in Medical Image Analysis: A Review

Information, Год журнала: 2025, Номер 16(3), С. 195 - 195

Опубликована: Март 3, 2025

Deep convolutional neural networks (CNNs) have revolutionized medical image analysis by enabling the automated learning of hierarchical features from complex imaging datasets. This review provides a focused CNN evolution and architectures as applied to analysis, highlighting their application performance in different fields, including oncology, neurology, cardiology, pulmonology, ophthalmology, dermatology, orthopedics. The paper also explores challenges specific outlines trends future research directions. aims serve valuable resource for researchers practitioners healthcare artificial intelligence.

Язык: Английский

---

A multi-scale convolutional LSTM-dense network for robust cardiac arrhythmia classification from ECG signals

Computers in Biology and Medicine, Год журнала: 2025, Номер 191, С. 110121 - 110121

Опубликована: Апрель 14, 2025

Cardiac arrhythmias are irregular heart rhythms that, if undetected, can lead to severe cardiovascular conditions. Detecting these anomalies early through electrocardiogram (ECG) signal analysis is critical for preventive healthcare and effective treatment. However, the automatic classification of poses significant challenges, including class imbalance noise interference in ECG signals. This paper introduces Multi-Scale Convolutional LSTM Dense Network (MS-CLDNet) model, an advanced deep-learning model specifically designed address issues improve arrhythmia accuracy other relevant metrics. aims develop efficient MS-CLDNet, accurately classifying cardiac from Addressing challenges like interference, integrates bidirectional long short-term memory (LSTM) networks temporal pattern recognition, Blocks feature refinement, Neural Networks (CNNs) robust extraction. To achieve accurate different types arrhythmias, Classification Head refines extracted features even further. Utilizing MIT-BIH dataset, key pre-processing techniques such as wavelet-based denoising were employed enhance clarity. Results indicate that MS-CLDNet achieves a 98.22 %, outperforming baseline models with low average loss values (0.084). research highlights how crucial it combine sophisticated neural network architectures precision automated diagnostic systems, which could have important applications detection.

Язык: Английский

---

Electrocardiogram Based Arrhythmia Classification Using Long Short-Term Memory with Luong Attention Mechanism

International journal of intelligent engineering and systems, Год журнала: 2024, Номер 17(3), С. 696 - 705

Опубликована: Май 3, 2024

The Electrocardiogram (ECG) serves as a crucial indicator of diverse cardiac conditions, emphasizing the importance precise signal classification for automated arrhythmia detection.ECG is an efficient tool diagnosis and detection arrhythmia.Detecting arrhythmias in extended ECG segments can result episodes being overlooked.However, since transmits massive amount information, it becomes very complex challenging to extract relevant information from visual analysis.To overcome this problem, research proposes Long Short-Term Memory (LSTM) with Luong Attention Mechanism approach into 5 classes.When LSTM combined attention, they learn which parts are at each time step, effectively capturing both short-term long-term dependencies.For evaluating performance proposed method, data collected benchmark dataset called MIT-BIH dataset.After collection dataset, pre-processing done using Continuous Wavelet Transform (CWT) reduce low high-frequency noise.After that, pre-processed forwarded feature extraction process features by statistical (Skewness, Kurtosis, Moment, etc.) time-frequency domain features.Finally, used classify classes.From analysis, achieved better results overall metrics.The method achieves accuracy 99.75% comparatively higher than existing approaches like Deep Residual Convolutional Neural Network (DRCNN) Depth wise Separable CNN Focal Loss (DSC-FL-CNN).

Язык: Английский

---

Classification of imbalanced ECGs through segmentation models and augmented by conditional diffusion model

PeerJ Computer Science, Год журнала: 2024, Номер 10, С. e2299 - e2299

Опубликована: Сен. 4, 2024

Electrocardiograms (ECGs) provide essential data for diagnosing arrhythmias, which can potentially cause serious health complications. Early detection through continuous monitoring is crucial timely intervention. The Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) arrhythmia dataset employed analysis research comprises imbalanced data. It necessary to create a robust model independent imbalances classify arrhythmias accurately. To mitigate the pronounced class imbalance in MIT-BIH dataset, this study employs advanced augmentation techniques, specifically variational autoencoder (VAE) and conditional diffusion, augment dataset. Furthermore, accurately segmenting heartbeat into individual heartbeats confidently detecting arrhythmias. This compared that annotation-based segmentation, utilizing R-peak labels, utilized an automated segmentation method based on deep learning segment heartbeats. In our experiments, proposed model, MobileNetV2 along with diffusion address minority class, demonstrated notable 1.23% improvement F1 score 1.73% precision, classifying classes original presents classifies wide range including classes, moving beyond previously limited classification models. serve as basis better utilization performance diagnosis medical service research. These achievements enhance applicability field contribute improving quality healthcare services by providing more sophisticated reliable diagnostic tools.

Язык: Английский

---

Low-cost electrocardiogram monitoring system for elderly people using LabVIEW

International Journal of Reconfigurable and Embedded Systems (IJRES), Год журнала: 2024, Номер 13(2), С. 483 - 483

Опубликована: Март 26, 2024

Cardiovascular diseases increase due to factors such as obesity, an inadequate diet, and are a problem shortages of medical personnel hospitals. In this case, the implementation technological solutions is presented necessity prevent heart diseases. Various approaches used design low-cost electrocardiogram (ECG) devices, from use Bluetooth technology facilitate data transmission, development wearable ECG devices that artificial intelligence. The objective develop monitoring system in LabVIEW visualize rhythms older adults city Lima (Peru), focusing on ease adaptation their needs, with purpose collaboration between health professionals. A approach encompasses design, implementation, iterative testing, well practical evaluations pilot testing. As result, correct functioning device was validated. Electronic components electrodes were integrated into board capture cardiac signals, energized batteries sending information interface LabVIEW. conclusion, portable has been developed uses operational amplifiers (Op-amps) analog filters reduce noise measurements intuitive

Язык: Английский

---

Multi-input Deep Learning Model for RP Diagnosis Using FVEP and Prior Knowledge

Lecture notes in computer science, Год журнала: 2024, Номер unknown, С. 287 - 299

Опубликована: Янв. 1, 2024

Язык: Английский

---

Cardiac Arrhythmia Classification Using Convolutional Neural Network

Опубликована: Янв. 1, 2024

Язык: Английский

---

Cost and Renewable Energy Management by IoT‐Oriented Smart Home Based on Smart Grid Demand Response

Опубликована: Ноя. 17, 2024

The Internet has become an integral aspect of human lives, enabling the remote monitoring and management various equipment such as televisions, air conditioners, refrigerators, washing machines. This enhanced functionality is made possible by implementing things (IoT) technology, which imbues these items with more intelligence. Smart Home apps, a constituent intelligent cities, undoubtedly represent one most sought-after applications. user's text needs to be longer rewritten academically. research presents design smart energy (SEM) system that utilizes NodeMCU Android platforms. SEM specifically developed component home application. enables real-time use, along capability capture data about device operating times consumption statistics. Furthermore, optimizes use meeting maximum requirements during reduced prices. interface allows customers monitor modify their power usage patterns, aiming enhance efficiency. Additionally, it gives create daily weekly schedules.

Язык: Английский

---