Research Output
Articles
Alamatsaz, N.,
Tabatabaei, L.,
Yazdchi, M.,
Payan, H.,
Alamatsaz, N.,
Nasimi, F. Publication Date: 2024
Biomedical Signal Processing and Control (17468108)90
Objective: Electrocardiogram (ECG) is the most frequent and routine diagnostic tool used for monitoring heart electrical signals and evaluating its functionality. The human heart can suffer from a variety of diseases, including cardiac arrhythmias. Arrhythmia is an irregular heart rhythm that in severe cases can lead to stroke and can be diagnosed via ECG recordings. Since early detection of cardiac arrhythmias is of great importance, computerized and automated classification and identification of these abnormal heart signals have received much attention for the past decades. Methods: This paper introduces a light Deep Learning (DL) approach for high accuracy detection of 8 different cardiac arrhythmias and normal rhythms. To employ DL techniques, the ECG signals were preprocessed using resampling and baseline wander removal techniques. The classification was performed using an 11-layer network employing a combination of Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM). Results: In order to evaluate the proposed technique, ECG signals are chosen from the two physionet databases, the MIT-BIH arrhythmia database and the long-term AF database. The proposed DL framework based on the combination of CNN and LSTM showed promising results than most of the state-of-the-art methods. The proposed method reaches the mean diagnostic accuracy of 98.24%. Conclusion: A trained model for arrhythmia classification using diverse ECG signals were successfully developed and tested. Significance: This study presents a lightweight classification technique with high diagnostic accuracy compared to other notable methods, making it a potential candidate for implementation in Holter monitor devices for arrhythmia detection. Finally, we used SHapley Additive exPlanations (SHAP), the most popular Explainable Artificial Intelligence (XAI) method to understand how our model make predictions. The results indicate that those features (ECG samples) that have contributed the most to a prediction are consonant with clinicians’ decisions. Therefore, the use of interpretable models increases the trust of clinicians in AI and thus leads to decreasing the number of misdiagnoses of cardiovascular diseases. © 2023 Elsevier Ltd
Publication Date: 2022
Biomedical Signal Processing and Control (17468108)78
Proposing a practical method for high-performance emotion recognition could facilitate human–computer interaction. Among existing methods, deep learning techniques have improved the performance of emotion recognition systems. In this work, a new multimodal neural design is presented wherein audio and visual data are combined as the input to a hybrid network comprised of a bidirectional long short term memory (BiLSTM) network and two convolutional neural networks (CNNs). The spatial and temporal features extracted from video frames are fused with Mel-Frequency Cepstral Coefficients (MFCCs) and energy features extracted from audio signals and BiLSTM network outputs. Finally, a Softmax classifier is used to classify inputs into the set of target categories. The proposed model is evaluated on Surrey Audio–Visual Expressed Emotion (SAVEE), Ryerson Audio–Visual Database of Emotional Speech and Song (RAVDESS), and Ryerson Multimedia research Lab (RML) databases. Experimental results on these datasets prove the effectiveness of the proposed model where it achieves the accuracy of 99.75%, 94.99%, and 99.23% for the SAVEE, RAVDESS, and RML databases, respectively. Our experimental study reveals that the suggested method is more effective than existing algorithms in adapting to emotion recognition in these datasets. © 2022 Elsevier Ltd
Publication Date: 2022
PLoS ONE (19326203)17(2 February)
Differentiating between shockable and non-shockable Electrocardiogram (ECG) signals would increase the success of resuscitation by the Automated External Defibrillators (AED). In this study, a Deep Neural Network (DNN) algorithm is used to distinguish 1.4-second segment shockable signals from non-shockable signals promptly. The proposed technique is frequency-independent and is trained with signals from diverse patients extracted from MIT-BIH, MIT-BIH Malignant Ventricular Ectopy Database (VFDB), and a database for ventricular tachyarrhythmia signals from Creighton University (CUDB) resulting, in an accuracy of 99.1%. Finally, the raspberry pi minicomputer is used to load the optimized version of the model on it. Testing the implemented model on the processor by unseen ECG signals resulted in an average latency of 0.845 seconds meeting the IEC 60601-2-4 requirements. According to the evaluated results, the proposed technique could be used by AED’s. Copyright: © 2022 Nasimi, Yazdchi. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Publication Date: 2022
PLoS ONE (19326203)17(2 February)
The demand for long-term continuous care has led healthcare experts to focus on development challenges. On-chip energy consumption as a key challenge can be addressed by data reduction techniques. In this paper, the pseudo periodic nature of ElectroCardioGram (ECG) signals has been used to completely remove redundancy from frames. Compressing aligned QRS complexes by Compressed Sensing (CS), result in highly redundant measurement vectors. By removing this redundancy, a high cluster of near zero samples is gained. The efficiency of the proposed algorithm is assessed using the standard MIT-BIH database. The results indicate that by aligning ECG frames, the proposed technique can achieve superior reconstruction quality compared to state-of-the-art techniques for all compression ratios. This study proves that by aligning ECG frames with a 0.05% unaligned frame rate(R-peak detection error), more compression could be gained for PRD > 5% when 5-bit non-uniform quantizer is used. Furthermore, analysis done on power consumption of the proposed technique, indicates that a very good recovery performance can be gained by only consuming 4.9μW more energy per frame compared to traditional CS. Copyright: © 2022 Nasimi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Publication Date: 2020
Biomedical Signal Processing and Control (17468108)60
Background and objectives: Data compression techniques have been used in order to reduce power consumption when transmitting electrocardiogram (ECG) signals in wireless body area networks (WBAN). Among these techniques, compressed sensing allows sparse or compressible signals to be encoded with only a small number of measurements. Although ECG signals are not sparse, they can be made sparse in another domain. Numerous sparsifying techniques are available, but when signal quality and energy consumption are important, existing techniques leave room for improvements. Methods: To leverage compressed sensing, we increased the sparsity of an ECG frame by removing the redundancy in a normal frame. In this study, by framing a signal according to the detected QRS complex (R peaks), consecutive frames of the signal become highly similar. This helps remove redundancy and consequently makes each frame sparse. In order to increase detection performance, different frames that symptomize a cardiovascular disease are sent uncompressed. Results: For evaluating and comparing our proposed technique with different state-of-the-art techniques two datasets that contained normal and abnormal ECG: MIT-BIH Arrhythmia Database and MIT-BIH Long Term Database were used. For performance evaluation, we performed heart rate variability (HRV) analysis as well as energy-based distortion analysis. The proposed method reaches an accuracy of 99.9%, for a compression ratio of 25. For MIT-BIH Long Term Database, the average percentage root-mean squared difference (PRD) is less than 10 for all compression ratios. Conclusion: Removing the redundancy between successive similar frames and exact transmission of dissimilar frames, the proposed method proves to be appropriate for heart rate variability analysis and abnormality detection. © 2020
