RESUMEN
The scarcity of annotated data is a common issue in the realm of heartbeat classification based on deep learning. Transfer learning (TL) has emerged as an effective strategy for addressing this issue. However, current TL techniques in this realm overlook the probability distribution differences between the source domain (SD) and target domain (TD) databases. The motivation of this paper is to address the challenge of labeled data scarcity at the model level while exploring an effective method to eliminate domain discrepancy between SD and TD databases, especially when SD and TD are derived from inconsistent tasks. This study proposes a multi-module heartbeat classification algorithm. Initially, unsupervised feature extractors are designed to extract rich features from unlabeled SD and TD data. Subsequently, a novel adaptive transfer method is proposed to effectively eliminate domain discrepancy between features of SD for pre-training (PTF-SD) and features of TD for fine-tuning (FTF-TD). Finally, the adapted PTF-SD is employed to pre-train a designed classifier, and FTF-TD is used for classifier fine-tuning, with the objective of evaluating the algorithm's performance on the TD task. In our experiments, MNIST-DB serves as the SD database for handwritten digit image classification task, MIT-DB as the TD database for heartbeat classification task. The overall accuracy of classifying heartbeats into normal heartbeats, supraventricular ectopic beats (SVEBs), and ventricular ectopic beats (VEBs) reaches 96.7 %. Specifically, the sensitivity (Sen), positive predictive value (PPV), and F1 score for SVEBs are 0.802, 0.701, and 0.748, respectively. For VEBs, Sen, PPV, and F1 score are 0.976, 0.840, and 0.903, respectively. The results indicate that the proposed multi-module algorithm effectively addresses the challenge labeled data scarcity in heartbeat classification through unsupervised learning and adaptive feature transfer methods.
Asunto(s)
Aprendizaje Automático no Supervisado , Complejos Prematuros Ventriculares , Humanos , Frecuencia Cardíaca , Electrocardiografía/métodos , Procesamiento de Señales Asistido por Computador , AlgoritmosRESUMEN
OBJECTIVE: The proliferation of wearable devices has escalated the standards for photoplethysmography (PPG) signal quality. This study introduces a lightweight model to address the imperative need for precise, real-time evaluation of PPG signal quality, followed by its deployment and validation utilizing our integrated upper computer and hardware system. METHODS: Multiscale Markov Transition Fields (MMTF) are employed to enrich the morphological information of the signals, serving as the input for our proposed hybrid model (HM). HM undergoes initial pre-training utilizing the MIMIC-III and UCI databases, followed by fine-tuning the Queensland dataset. Knowledge distillation (KD) then transfers the large-parameter model's knowledge to the lightweight hybrid model (LHM). LHM is subsequently deployed on the upper computer for real-time signal quality assessment. RESULTS: HM achieves impressive accuracies of 99.1% and 96.0% for binary and ternary classification, surpassing current state-of-the-art methods. LHM, with only 0.2 M parameters (0.44% of HM), maintains high accuracy despite a 2.6% drop. It achieves an inference speed of 0.023 s per image, meeting real-time display requirements. Furthermore, LHM attains a 97.7% accuracy on a self-created database. HM outperforms current methods in PPG signal quality accuracy, demonstrating the effectiveness of our approach. Additionally, LHM substantially reduces parameter count while maintaining high accuracy, enhancing efficiency and practicality for real-time applications. CONCLUSION: The proposed methodology demonstrates the capability to achieve high-precision and real-time assessment of PPG signal quality, and its practical validation has been successfully conducted during deployment. SIGNIFICANCE: This study contributes a convenient and accurate solution for the real-time evaluation of PPG signals, offering extensive application potential.
Asunto(s)
Procesamiento de Señales Asistido por Computador , Dispositivos Electrónicos Vestibles , Humanos , Algoritmos , Fotopletismografía/métodos , Frecuencia Cardíaca , ArtefactosRESUMEN
OBJECTIVE: Obstructive sleep apnea (OSA) is a common sleep-related breathing disorder that can lead to a wide range of health issues if left untreated. This study aims to address the lack of research on personalized models for single-lead electrocardiogram (ECG)-based OSA detection, by proposing an automatic semi-supervised algorithm for automated low-cost personalization fine-tuning. METHODS: We utilize a convolutional neural network (CNN)-based auto-encoder (AE) with a modified training objective to detect anomalous region of OSA. An indicator based on model outputs is utilized as a benchmark measure to assign pseudo-labels with confidence to each sample. Finally, we perform validation of the semi-supervised algorithm on the same database and cross-database scenarios. RESULTS: By introducing semi-supervised personalization, the accuracy, AUC, and mean absolute error (MAE) of the general model (GM) of 35 subjects from the same database are improved from 86.3%, 0.915, and 5.178 to 90.3%, 0.948, and 2.593. Simultaneously, in the validation of 25 subjects from a cross-database, the accuracy, AUC, and MAE of the GM are enhanced from 75.6%, 0.800, and 9.149 to 84.3%, 0.881, and 3.509. CONCLUSION: The improved version of AE demonstrates excellent adaptability in identifying abnormal features in OSA, employing a data-driven approach to assign pseudo-labels for unknown data automatically. Additionally, leveraging the pseudo-labels through a semi-supervised fine-tuning strategy provides a solution to overcome the limitation of clinical annotations, facilitating low-cost implementation of personalized models. SIGNIFICANCE: The semi-supervised approach proposed in this article provides a high-performance and annotation-free solution for personalized adjustment of automatic OSA detection.
Asunto(s)
Aprendizaje Profundo , Apnea Obstructiva del Sueño , Humanos , Apnea Obstructiva del Sueño/diagnóstico , Respiración , Aprendizaje Automático Supervisado , ElectrocardiografíaRESUMEN
OBJECTIVE: Epilepsy prediction algorithms offer patients with drug-resistant epilepsy a way to reduce unintended harm from sudden seizures. The purpose of this study is to investigate the applicability of transfer learning (TL) technique and model inputs for different deep learning (DL) model structures, which may provide a reference for researchers to design algorithms. Moreover, we also attempt to provide a novel and precise Transformer-based algorithm. METHODS: Two classical feature engineering methods and the proposed method which consists of various EEG rhythms are explored, then a hybrid Transformer model is designed to evaluate the advantages over pure convolutional neural networks (CNN)-based models. Finally, the performances of two model structures are analyzed utilizing patient-independent approach and two TL strategies. RESULTS: We tested our method on the CHB-MIT scalp EEG database, the results showed that our feature engineering method gains a significant improvement in model performance and is more suitable for Transformer-based model. In addition, the performance improvement of Transformer-based model utilizing fine-tuning strategies is more robust than that of pure CNN-based model, and our model achieved an optimal sensitivity of 91.7% with false positive rate (FPR) of 0.00/h. CONCLUSION: Our epilepsy prediction method achieves excellent performance and demonstrates its advantage over pure CNN-based structure in TL. Moreover, we find that the information contained in the gamma ( γ ) rhythm is helpful for epilepsy prediction. SIGNIFICANCE: We propose a precise hybrid Transformer model for epilepsy prediction. The applicability of TL and model inputs is also explored for customizing personalized models in clinical application scenarios.
Asunto(s)
Epilepsia , Humanos , Epilepsia/diagnóstico , Convulsiones/diagnóstico , Redes Neurales de la Computación , Electroencefalografía/métodos , Algoritmos , Aprendizaje AutomáticoRESUMEN
Objective. Automatic electrocardiogram (ECG) interpretation based on deep learning methods is attracting increasing attention. In this study, we propose a novel method to accurately classify multi-lead ECGs using deep residual neural networks.Approach. ECG recordings from seven different open databases were provided by PhysioNet/Computing in Cardiology Challenge 2021. All the ECGs were pre-processed to obtain the same sampling rate. The label inconsistency among the databases was corrected by adding or removing specific labels. A label mask was created to filter out potentially incorrectly labelled data. Five models based on deep residual convolutional neural networks were optimized using an asymmetric loss function to classify multi-lead ECGs.Main results. The proposed method achieved an official challenge score of 0.54, 0.52, 0.50, 0.51, and 0.50 on twelve-lead, six-lead, four-lead, three-lead, and two-lead ECG test sets, respectively. These scores were ranked 5th, 3rd, 7th, 5th and 7th, respectively, in the challenge.Significance. The proposed method can correct the differential labeling tendency of databases from different sources and exhibits good generalization for classifying multi-lead ECGs in the hidden test set. The proposed models have the potential for clinical applications.
Asunto(s)
Electrocardiografía , Redes Neurales de la Computación , Algoritmos , Bases de Datos Factuales , Progresión de la Enfermedad , Electrocardiografía/métodos , HumanosRESUMEN
Objective. Electrocardiography is a common method for screening cardiovascular diseases. Accurate heartbeat classification assists in diagnosis and has attracted great attention. In this paper, we proposed an automatic heartbeat classification method based on a transformer neural network using a self-attention mechanism.Approach.An adaptive heartbeat segmentation method was designed to selectively focus on the time-dependent representation of heartbeats. A one-dimensional convolution layer was used to embed wave characteristics into symbolic representations, and then, a transformer block using multi-head attention was applied to deal with the dependence of wave-embedding. The model was trained and evaluated using the MIT-BIH arrhythmia database (MIT-DB). To improve the model performance, the model pre-trained on MIT-BIH supraventricular arrhythmia database (MIT-SVDB) was used and fine-tuned on MIT-DB.Main results.The proposed method was verified using the MIT-DB for two groups. In the first group, our method attained F1 scores of 0.86 and 0.96 for the supraventricular ectopic beat class and ventricular ectopic beat class, respectively. In the second group, our method achieved an average F1 value of 99.83% and better results than other state-of-the-art methods.Significance.We proposed a novel heartbeat classification method based on a transformer model. This method provides a new solution for real-time electrocardiogram heartbeat classification, which can be applied to wearable devices.