Revised Tunable Q-Factor Wavelet Transform for EEG-Based Epileptic Seizure Detection

A new VCG signal compression technique based on discrete Karhunen-Loeve expansion and tunable quality wavelet transform

This paper presents a novel two-stage compression technique for vectorcardiogram (VCG) signals, combining Discrete Karhunen-Loeve (K-L) Expansion and Tunable Q-Factor Wavelet Transform (TQWT). In the first stage, VCG signals undergo discrete K-L expansion and a secondary rotation to reduce variance caused by physiological factors, such as respiration and varying heart orientations among patients. This process effectively simplifies the dataset by leveraging eigenvector-based transformation, while standardizing the data across different VCG records. In the second stage, the standardized data is processed using TQWT, with finely tuned parameters (Q-factor of 4, redundancy factor of 1.2, and 6 stages), followed by quantization and Run-Length Encoding (RLE). The RLE method efficiently compresses the long sequences of zeros generated during the process, further enhancing the data reduction. The proposed method was rigorously evaluated using the PTB Diagnostic ECG Database, demonstrating remarkable compression efficiency. When compared with standard approaches like Discrete K-L Transform and Discrete Cosine Transform (DCT), the method achieved a superior average compression ratio of 15.43. Key evaluation metrics further highlight its efficacy, including an average Percent Root Difference (PRD) of 7.39 %, Fidelity of 99.72 %, Peak Signal-to-Noise Ratio (PSNR) of 37.38 dB, and a Quality Score (QS) of 2.13 %. Moreover, the method's rapid processing speed of 0.076 s per record makes it well-suited for real-time applications. This innovative approach provides an effective solution for VCG signal compression, enhancing the storage and transmission efficiency, while preserving high signal fidelity for clinical use.

Detection of Anxiety-Based Epileptic Seizures in EEG Signals Using Fuzzy Features and Parrot Optimization-Tuned LSTM

In humans, epilepsy is diagnosed through electroencephalography (EEG) signals. Epileptic seizures (ESs) arise due to anxiety. The detection of anxiety-based seizures is challenging for radiologists, and there is a limited availability of anxiety-based EEG signals. Data augmentation methods are required to increase the number of novel samples. An epileptic seizure arises due to anxiety, which manifests as variations in EEG signal patterns consisting of changes in the size and shape of the signal. In this study, anxiety EEG signals were synthesized by applying data augmentation methods such as random data augmentation (RDA) to existing epileptic seizure signals from the Bonn EEG dataset. The data-augmented anxiety seizure signals were processed using three algorithms-(i) fuzzy C-means-particle swarm optimization-long short-term memory (FCM-PS-LSTM), (ii) particle swarm optimization-long short-term memory (PS-LSTM), and (iii) parrot optimization LSTM (PO-LSTM)-for the detection of anxiety ESs via EEG signals. The predicted accuracies of detecting ESs through EEG signals using the proposed algorithms-namely, (i) FCM-PS-LSTM, (ii) PS-LSTM, and (iii) PO-LSTM-were about 98%, 98.5%, and 96%, respectively.

Positional multi-length and mutual-attention network for epileptic seizure classification

The automatic classification of epilepsy electroencephalogram (EEG) signals plays a crucial role in diagnosing neurological diseases. Although promising results have been achieved by deep learning methods in this task, capturing the minute abnormal characteristics, contextual information, and long dependencies of EEG signals remains a challenge. To address this challenge, a positional multi-length and mutual-attention (PMM) network is proposed for the automatic classification of epilepsy EEG signals. The PMM network incorporates a positional feature encoding process that extracts minute abnormal characteristics from the EEG signal and utilizes a multi-length feature learning process with a hierarchy residual dilated LSTM (RDLSTM) to capture long contextual dependencies. Furthermore, a mutual-attention feature reinforcement process is employed to learn the global and relative feature dependencies and enhance the discriminative abilities of the network. To validate the effectiveness PMM network, we conduct extensive experiments on the public dataset and the experimental results demonstrate the superior performance of the PMM network compared to state-of-the-art methods.

A novel finite spectral entropy: Gated term memory unit recursive network integrated with Ladybug Beetle Optimization algorithm for epileptic seizure detection

Professional medical experts use a visual electroencephalography (EEG) signal for epileptic seizure detection, although this method is time-consuming and highly subject to bias. The majority of previous epileptic detection techniques have poor efficiency, detection performance and also which are unsuited to handle large datasets. In order to solve the aforementioned issues and to assist medical professionals with an advanced technology, a computerized epileptic seizure detection system is essential. Therefore, the proposed work intends to design an automated detection tool for predicting an epileptic seizure from EEG signals. For this purpose, a novel non-linear feature analysis and deep learning algorithms are deployed in this work. Initially, the signal decomposition, filtering and artifacts removal operations are carried out with the use of finite Haar wavelet transformation technique. After that, the finite spectral entropy (FSE) based feature extraction model has been used to extract the time, frequency, and time-frequency features from the normalized signal. Consequently, the novel gated term memory unit recursive network (GTRN) model is employed to predict the given EEG signal as whether healthy or seizure affected including the class with high accuracy. During this process, the recently developed Ladybug Beetle Optimization (LBO) algorithm is used to compute the logistic sigmoid function based on the solution. The purpose of using this algorithm is to simplify the process of classification with increased seizure prediction accuracy and performance. Moreover, the standard and popular benchmark EEG datasets are used to validate and test the results of the proposed FSE-GTRN-LBO mechanism. By leveraging the finite Haar wavelet transformation and FSE-based feature extraction, we can efficiently process EEG signals. The utilization of the GTRN model enables accurate classification of healthy and seizure-affected EEG data. To optimize the classification process further, we integrate the LBO algorithm, streamlining the computation of the logistic sigmoid function. Through comprehensive validation on standard EEG datasets, our proposed FSE-GTRN-LBO mechanism achieves outstanding seizure prediction accuracy and performance, surpassing existing state-of-the-art techniques.