Compression of neonatal EEG seizure signalswith finite rate of innovation

Analyses of neonatal EEG seizures and subsequent diagnoses can only be done effectively on long-term recordings on the condition that the morphology of the EEG signals are retained. Therefore, a reliable, accurate and efficient compression and reconstruction technique is necessary to store and retrieve the data. In this paper, we propose a new compression technique for neonatal EEG seizure signals via sampling theory developed for signals with a finite rate of innovation. Firstly, the EEG seizure signals are modeled as periodic nonuniform linear splines. Next, through the sampling and reconstruction scheme developed for signals with finite rate of innovation, we show that neonatal EEG seizure signals can be highly compressed while preserving their morphologies.