The paper provides an overview of the removal of noise cancellation in ECG signals using an LMS filter in a system generator for monitoring ECG parameters and the study of the P wave to diagnose cardiac arrhythmia. The real ECG signals were evaluated from MIT-BIH database. Using Xilinx system Generator., the LMS adaptive filters technique is implemented. In order to efficiently verify the algorithm., the simulation of the models was carried out in MATLAB and Simulink. The core LMS adaptive filter and its fundamental basic building blocks technique was implemented in Xilinx System Generator. Here., high-pass least-square linear phase Finite Impulse Response (FIR) filtering approach to remove the baseline wander noise from the system's input ECG signal. A digital filter used in adaptive filtering has weights that are managed using adaptive algorithm to reduce difference between output of the filter and a reference signal that matches and fulfills the criterion. The reference signal's characteristics depends on the application under consideration. Convergence rate and steady state mean square error are the two main measures to evaluate the efficiency and performance of an adaptive filter. 

Noise Removal from ECG Signal using LMS Adaptive Filter Implementation in Xilinx System Generator

This paper proposes ReAdapt–a reconfigurable datapath architecture for scaling the energy-quality trade-off of adaptive filtering at runtime. The ReAdapt can dynamically select four adaptive filtering algorithms for gradating complexity levels during runtime by reconfiguring the processing flow in its datapath and by blocking the switching activity (e.g., reducing the CMOS dynamic power) of unused modules with data-gating. The ReAdapt proposal can scale the energy-quality trade-off by choosing the following four different levels of filter algorithms complexity: 1) least mean square (LMS); 2) partial update normalized LMS (PU-NLMS); 3) set-membership normalized LMS (SM-NLMS); 4) normalized LMS (NLMS). The ReAdapt architecture reuses common modules of each adaptive filter, resulting in a compact VLSI hardware implementation. The ReAdapt architecture operation is implemented in a case-study for interference mitigation for electroencephalogram (EEG) signal processing. The hardware synthesis results show an increase of 6.80 times in throughput and at least a reduction of 2.84 times in energy per operation compared with the state-of-the-art adaptive filters. This paper also investigates the benefits of dynamically reconfiguring the four ReAdapt operating modes at runtime for different levels of signal-to-noise ratio (SNR) for the processed signals. We also demonstrate that dynamically reconfiguring the ReAdapt operating modes during runtime results in an optimal energy-quality trade-off which is advantageous over the conventional single static mode. 

ReAdapt: A Reconfigurable Datapath for Runtime Energy-Quality Scalable Adaptive Filters

The paper provides an overview of the removal of noise cancellation in ECG signals using an LMS filter in a system generator for monitoring ECG parameters and the study of the P wave to diagnose cardiac arrhythmia. The real ECG signals were evaluated from MIT-BIH database. Using Xilinx system Generator., the LMS adaptive filters technique is implemented. In order to efficiently verify the algorithm., the simulation of the models was carried out in MATLAB and Simulink. The core LMS adaptive filter and its fundamental basic building blocks technique was implemented in Xilinx System Generator. Here., high-pass least-square linear phase Finite Impulse Response (FIR) filtering approach to remove the baseline wander noise from the system's input ECG signal. A digital filter used in adaptive filtering has weights that are managed using adaptive algorithm to reduce difference between output of the filter and a reference signal that matches and fulfills the criterion. The reference signal's characteristics depends on the application under consideration. Convergence rate and steady state mean square error are the two main measures to evaluate the efficiency and performance of an adaptive filter.

In this study, a multi-channel Electroencephalogram (EEG) mental fatigue detection algorithm is proposed based on the Convolutional Neural Network- Long Short-Term Memory (CNN-LSTM) network. The CNN-LSTM deep network model is used to distinguish three different mental fatigue states: awake, mild fatigue and severe fatigue. The model is validated on a self-collected dataset collected by the 2-BACK experimental paradigm. The data set contains fatigue data for a total of 10 healthy adults without adverse habits. The average recognition accuracy of the model is 97.12%. The model yields a sensitivity of 97.80% and a specificity of 99.28%. Results show that our proposed CNN-LSTM model can distinguish the three different mental fatigue states effectively. 

EEG-Based Mental Fatigue Detection Using CNN-LSTM

This paper proposes ReAdapt–a reconfigurable datapath architecture for scaling the energy-quality trade-off of adaptive filtering at runtime. The ReAdapt can dynamically select four adaptive filtering algorithms for gradating complexity levels during runtime by reconfiguring the processing flow in its datapath and by blocking the switching activity (e.g., reducing the CMOS dynamic power) of unused modules with data-gating. The ReAdapt proposal can scale the energy-quality trade-off by choosing the following four different levels of filter algorithms complexity: 1) least mean square (LMS); 2) partial update normalized LMS (PU-NLMS); 3) set-membership normalized LMS (SM-NLMS); 4) normalized LMS (NLMS). The ReAdapt architecture reuses common modules of each adaptive filter, resulting in a compact VLSI hardware implementation. The ReAdapt architecture operation is implemented in a case-study for interference mitigation for electroencephalogram (EEG) signal processing. The hardware synthesis results show an increase of 6.80 times in throughput and at least a reduction of 2.84 times in energy per operation compared with the state-of-the-art adaptive filters. This paper also investigates the benefits of dynamically reconfiguring the four ReAdapt operating modes at runtime for different levels of signal-to-noise ratio (SNR) for the processed signals. We also demonstrate that dynamically reconfiguring the ReAdapt operating modes during runtime results in an optimal energy-quality trade-off which is advantageous over the conventional single static mode.

Electroencephalography (EEG) provides diagnostic information related to various brain disorders. Various types of interferences, like line interference, EOG, and ECG, muscle movement, cause artifacts in EEG data. Therefore, denoising EEG data plays a vital role in preserving the specific frequency content of the signal. Several filtering techniques are available to detach the noise to preserve the integrity of EEG signals. In this paper, we have compared different filtering techniques i.e., Adaptive filters, LPF Butterworth filter, Notch filter, wavelets on epileptic EEG signals, and sleep EEG signal. Our result suggests that the wavelet transform is the best option for denoising the EEG signal as it is more efficient in denoising the EEG signal without losing the original information. To select the best suitable wavelet function for denoising, Symlet4, Haar, Daubechies4, Biorthogonal2.6, Coiflets3, Discrete Meyer, Reverse Biorthogonal 6.8, Reverse Biorthogonal 2.8 has been used, and it is observed that wavelet function Bio-orthogonal 2.6 is the best suitable for denoising of EEG signal. Finally, a comparison between different filters has been done by two parameters MSE, PSNR. After a comparative analysis, we conclude that a wavelet transform is a useful tool than other filtering techniques in noise removal while sustaining diagnostic information in both the signal.

Anjali W. Pise

Papers

Comparative Analysis of Various Filtering Techniques for Denoising EEG Signals