Detection of epileptic events using genetic programming
Bruno Marchesi, A.L. Stelle, Heitor Silvério Lopes · 2002
This paper presents a method using genetic programming for automatic detection of 3 Hz spike-and-slow-wave complexes, that are a characteristic of typical absences, in electroencephalogram (EEG) signals. Training features are extracted from 1s EEG frames, randomly chosen from pre-recorded files. The frames are visually classified as spike-and-slow-wave complexes (SASWC) or non-spike-and-slow-wave complexes (NSASWC). Genetic programming techniques are then applied to these data to build a program capable of recognizing such complexes.