Implementation of EEG feature extractor and classifier for seizure detection on FPGA

Seizure is a neurological disorder that affects the brain and changes the activity of the brain. The detection and analysis of seizures is very important. Many different automatic methods are being searched by the experts for detecting the disease. Here we use a Field Programmable Gate Array(FPGA) based classifier to detect seizures in patients. Electroencephalography(EEG) is the best signal that can be used for signal analysis. The EEG signal is filtered using FIR filter to smoothen the signal. Then the features are extracted from the signal using Discrete Wavelet Transform(DWT). The filtered EEG data is processed by a Linear classifier, to detect the seizure in EEG data. The design was developed in Verilog HDL, simulated and implemented on Xilinx FPGA. FIR filter coefficients had been found out by using FDA tool command on MATLAB. It was found that the proposed system was able to classify normal and seizure signals with high accuracy, sensitivity and specificity as compared with previous systems.