International Journal of Engineering and Technology

Epileptic seizure detection in a large database of Electroencephalography (EEG) signals needs to be a time constrained process for real-time analysis. Epileptic seizure detection algorithms are designed to obtain and analyze a group of neural signals and recognize the presence of seizure occurrence. The computational cost of the algorithms should be minimized to reduce the processing time and memory consumption. Automated epileptic seizure detection using optimized feature selection improves the classification accuracy, but it occupies more processing time during the Artifact Removal (AR) stage. So, the execution time is greatly reduced by introducing task parallelism in the artifact removal stage. By harnessing parallel computing the computational overhead and processing time are decreased. An epileptic seizure detection application is developed and analyzed with respect to execution time, speedup, and parallel efficiency. The application was developed in Intel Pentium(R) Dual-core CPU with processor clock rate of 2.60 GHz, memory of 1.96 GB, and operating system of Windows XP Professional Service Pack 2.