Spike Detection from Electroencephalogram Signals with Aid of Hybrid Genetic Algorithm-Particle Swarm Optimization

In the clinical diagnosis of epilepsy, Spike detection in EEG is important. Many techniques have been proposed to detect the spikes present in EEG. In the previous spike detection method, the spikes were detected from the EEG signals using the morphological and hybrid optimization methods. It yielded good precision rate and hit rate but the detection rate is not in an adequate level. So we proposed a new optimized amplitude threshold technique to detect the spikes on the EEG signals. Initially, the EEG signals are preprocessed using Haar Wavelet to remove the noise. Then the Preprocessed signals are passed through the morphological filter to detect the spikes. The morphological filter utilizes hybrid GAPSO (Genetic Algorithm-Particle Swarm Optimization) technique to select the optimal structuring element. In order to improve the number of correctly detected spikes, we are utilized an optimized amplitude threshold technique. The optimized amplitude threshold technique is exploited Improved Particle swarm Optimization (IPSO) which is based on fractional velocity. Our proposed method is implemented on MATLAB platform and it is evaluated by using a set of EEG signals. The performance of our proposed technique is compared with the existing spike detection and optimization methods GA, PSO.