A novel approach for suppression of powerline interference and impulse noise in ECG signals

Author(s):  
Vikrant Bhateja ◽  
Shabana Urooj ◽  
Rishendra Verma ◽  
Rini Mehrotra
2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Aline dos Santos Silva ◽  
Hugo Almeida ◽  
Hugo Plácido da Silva ◽  
António Oliveira

AbstractMultiple wearable devices for cardiovascular self-monitoring have been proposed over the years, with growing evidence showing their effectiveness in the detection of pathologies that would otherwise be unnoticed through standard routine exams. In particular, Electrocardiography (ECG) has been an important tool for such purpose. However, wearables have known limitations, chief among which are the need for a voluntary action so that the ECG trace can be taken, battery lifetime, and abandonment. To effectively address these, novel solutions are needed, which has recently paved the way for “invisible” (aka “off-the-person”) sensing approaches. In this article we describe the design and experimental evaluation of a system for invisible ECG monitoring at home. For this purpose, a new sensor design was proposed, novel materials have been explored, and a proof-of-concept data collection system was created in the form of a toilet seat, enabling ECG measurements as an extension of the regular use of sanitary facilities, without requiring body-worn devices. In order to evaluate the proposed approach, measurements were performed using our system and a gold standard equipment, involving 10 healthy subjects. For the acquisition of the ECG signals on the toilet seat, polymeric electrodes with different textures were produced and tested. According to the results obtained, some of the textures did not allow the acquisition of signals in all users. However, a pyramidal texture showed the best results in relation to heart rate and ECG waveform morphology. For a texture that has shown 0% signal loss, the mean heart rate difference between the reference and experimental device was − 1.778 ± 4.654 Beats per minute (BPM); in terms of ECG waveform, the best cases present a Pearson correlation coefficient above 0.99.


2020 ◽  
Vol 10 (10) ◽  
pp. 2259-2273
Author(s):  
M. Suresh Kumar ◽  
G. Krishnamoorthy ◽  
D. Vaithiyanathan

This paper presents an adaptive ECG enhancement procedure based on Synchrosqueezing Transform (SST) to eliminate Powerline interference (PLI) from ECG signal. This work also incorporates the principles of modified discrete cosine transform (MDCT) and wiener filter. PLI is a major source of artifacts in the ECG signal which can affect its interpretation. Separating PLI from ECG signal poses a great challenge in the ECG analysis. The existing PLI removal techniques suffer from two major drawbacks such as Mode Mixing, inability to deal with non-stationary characteristics of signal. In this paper, we propose SST based wiener filtering approaches which can overcome the limitation of existing PLI suppression techniques. The proposed approaches undergo three stages of operation: mode decomposition, mode determination and peak restoration to filter out PLI from ECG recording. The mode decomposition uses SST to decompose the corrupted ECG signal into a sum of well separated intrinsic mode functions (IMFs). The objective is to filter out PLI from these IMFs. To do so, mode determination step which is based on Kurtosis and Crest factor is applied to categorize decomposition result into groups such as signal mode and noisy mode. Direct subtraction of the noisy mode from the corrupted ECG observation results in ECG signal with reduced peak since noise mode carries part of signal components in addition to interference. Hence, to restore the peak, wiener filter is applied on noisy mode to estimate actual PLI component. Finally, Noise free ECG signal is reconstructed by subtracting estimated PLI from the corrupted ECG signal. This paper discusses four possible PLI suppression methods which are derived by combining SST domain with wiener filter in various ways. Simulations are carried out to test the effectiveness of proposed methods. It is evident from the simulation results that the proposed methods can remove PLI of 50 Hz and its harmonics. The proposed techniques effectively removed PLI in both real and artificial ECG signals and to test its performance they are compared with state of the art methods. The SST based filtering methods outperformed other methods under the condition of PLI frequency variations. The experimental results also suggest that the SST based wiener filtering with modified reference approach offers better PLI suppression than all other methods.


Author(s):  
Rashmi Kumari ◽  
Anupriya Asthana ◽  
Vikas Kumar

Restoration of digital images degraded by impulse noise is still a challenge for researchers. Various methods proposed in the literature suffer from common drawbacks: such as introduction of artifacts and blurring of the images. A novel idea is proposed in this paper where presence of impulsive pixels are detected by ANFIS (Adaptive Neuro-Fuzzy Inference System) and mean of the median of suitable window size of noisy image is taken for the removal of the detected corrupted pixels. Experimental results show the effectiveness of the proposed restoration method both by qualitative and quantitative analysis.


2020 ◽  
Vol 10 (6) ◽  
pp. 1265-1273
Author(s):  
Lili Chen ◽  
Huoyao Xu

Sleep apnea (SA) is a common sleep disorders affecting the sleep quality. Therefore the automatic SA detection has far-reaching implications for patients and physicians. In this paper, a novel approach is developed based on deep neural network (DNN) for automatic diagnosis SA. To this end, five features are extracted from electrocardiogram (ECG) signals through wavelet decomposition and sample entropy. The deep neural network is constructed by two-layer stacked sparse autoencoder (SSAE) network and one softmax layer. The softmax layer is added at the top of the SSAE network for diagnosing SA. Afterwards, the SSAE network can get more effective high-level features from raw features. The experimental results reveal that the performance of deep neural network can accomplish an accuracy of 96.66%, a sensitivity of 96.25%, and a specificity of 97%. In addition, the performance of deep neural network outperforms the comparison models including support vector machine (SVM), random forest (RF), and extreme learning machine (ELM). Finally, the experimental results reveal that the proposed method can be valid applied to automatic SA event detection.


Information ◽  
2021 ◽  
Vol 12 (2) ◽  
pp. 63
Author(s):  
Fatima Sajid Butt ◽  
Luigi La Blunda ◽  
Matthias F. Wagner ◽  
Jörg Schäfer ◽  
Inmaculada Medina-Bulo ◽  
...  

Fall is a prominent issue due to its severe consequences both physically and mentally. Fall detection and prevention is a critical area of research because it can help elderly people to depend less on caregivers and allow them to live and move more independently. Using electrocardiograms (ECG) signals independently for fall detection and activity classification is a novel approach used in this paper. An algorithm has been proposed which uses pre-trained convolutional neural networks AlexNet and GoogLeNet as a classifier between the fall and no fall scenarios using electrocardiogram signals. The ECGs for both falling and no falling cases were obtained as part of the study using eight volunteers. The signals are pre-processed using an elliptical filter for signal noises such as baseline wander and power-line interface. As feature extractors, frequency-time representations (scalograms) were obtained by applying a continuous wavelet transform on the filtered ECG signals. These scalograms were used as inputs to the neural network and a significant validation accuracy of 98.08% was achieved in the first model. The trained model is able to distinguish ECGs with a fall activity from an ECG with a no fall activity with an accuracy of 98.02%. For the verification of the robustness of the proposed algorithm, our experimental dataset was augmented by adding two different publicly available datasets to it. The second model can classify fall, daily activities and no activities with an accuracy of 98.44%. These models were developed by transfer learning from the domain of real images to the medical images. In comparison to traditional deep learning approaches, the transfer learning not only avoids “reinventing the wheel,” but also presents a lightweight solution to otherwise computationally heavy problems.


Author(s):  
J. Mateo ◽  
C. Sanchez ◽  
A. Tortes ◽  
R. Cervigon ◽  
J.J. Rieta

Sign in / Sign up

Export Citation Format

Share Document