The study aims to establish an FPGA design model for epileptic seizures with discrete wavelet decomposition (DWT) and principal component analysis (PCA) to determine the optimum parameters of support vector machine (SVMs) for the EEG classification data. The FPGA Hardware implementation is described in this paper. Firstly, an optimized software-based medical diagnostic approach has been developed to determine the EEG class using only the variance calculated for each DWT level. This features extracted optimization leads to reduce the FPGA prototype size and to save energy consumption. Secondly, the proposed method has been designed and implemented on the Nexys 4 Artix 7 board using the Xilinx System Generator (XSG) for DSP. The performance evaluation of the proposed system has been made through two comparative studies, the first one, between the floating-point Matlab results and the fixed-point XSG results. The classification performances obtained from the proposed FPGA fixed-point implementation were compared to those obtained from the MATLAB floating-point. The second comparison was performed between the resulting performances and those obtained with the existing work in literature.

FPGA implementation of Epileptic Seizure detection based on DWT, PCA and Support Vector Machine

Lower limb amputees suffer from mobility limitations that lead to a degradation of their quality of life. Recently, inertial measurement units have been widely used for the analysis and recognition of gait activities which are two essential works for the control of lower limb prostheses. In this paper, accelerometer signals from a dataset are used to recognize gait activities. Five lower limb movements are classified and recognized using linear discriminant analysis (LDA) for dimension reduction and support vector machine (SVM) based on a linear function kernel for classification, respectively. The results prove that classification based on time-domain features and accelerometer signals can achieve 95.69% recognition accuracy for gait activity recognition. 

Classification of gait patterns in the time domain using accelerometer signal

Nowadays, a computer-aided diagnosis system is required to monitor the cardiac patients continuously and detecting the heart diseases automatically. In this paper, a new field programmable gate array-based morphological feature extraction approach is proposed for electrocardiogram signal analysis. The proposed architecture is mainly based on the Generalized Synchrosqueezing transform but a detrended fluctuation analyzer is applied in the reconstruction stage for capturing the maximum information of QRS complexes and P-waves by eliminating a set of noisy intrinsic modes. Then, a correntropy envelope is determined from the QRS enhanced signal for localizing the QRS region accurately. Also, an adaptive heuristic framework is introduced to detect the true P-wave from the P-wave enhanced reconstructed signal by analyzing both the positive and negative amplitudes. In addition, a root mean square Error estimation-based adaptive thresholding approach is used to estimate the T-wave after removing the P-QRS complexes. The proposed architecture has been implemented on field programmable gate array using the Xilinx Vertex 7 platform. The performance of the proposed architecture is validated by performing a comparative study between the resultant performances and those attained with state-of-the-art feature descriptors, in terms of Sensitivity, accuracy, positive prediction, error rate and field programmable gate array resources estimation. The proposed sensitivity, accuracy and positive prediction are 99.84%, 99.85% and 99.86% for QRS detection approach. The proposed sensitivity, accuracy and positive prediction are 99.45%, 99.23% and 99.78% for P-wave detection approach. The proposed sensitivity, accuracy and positive prediction are 99.58%, 99.65% and 100% for T-wave detection approach. The simulation results show that the proposed architecture overtakes existing designs and minimizes hardware complexity, which proves the suitability of this approach on real-time applications of electrocardiogram signals. 

A Novel Morphological Feature Extraction Approach for ECG Signal Analysis Based on Generalized Synchrosqueezing Transform, Correntropy Function and Adaptive Heuristic Framework in FPGA

People with lower limb amputations suffer from mobility limitations that degrade their quality of life. In this paper, we propose a hardware system dedicated to human walking activity recognition for smart prostheses, which uses a support vector machine (SVM) algorithm and time domain features to perform activity classification. To achieve a flexible and efficient hardware design, the architecture is implemented on FPGA Nexys 4 Artix 7 board using the Xilinx System Generator (XSG) for DSP. The performance evaluation of the proposed system has been done through a comparative study, the comparison has been done between floating point MATLAB results and fixed point XSG results.

FPGA Implementation of Support Vector Machine for Gait Activity Classification

FPGA - Based ECG signal analysis for arrhythmia detection system using SVM classifier

Due to the rising number of cardiovascular diseases death, the monitoring of cardiac patients has become a primary objective in the world. In this context, a fully FPGA-based system, for ECG signal...

Mohammed A. Ait

Papers

Fpga implementation system for qrs complex detection