Performance analysis of a 2-D EEG compression algorithm using an automatic seizure detection system

A recently developed compression algorithm that uses DWT, SPIHT and smoothness transforms to compress EEG channels in 2-D proved to give very low distortion values for high compression ratios. Although RD performance is a commonly used metric in signal compression, in medical signals, it is important to preserve important diagnostic information. In order to move towards such a diagnostics-oriented performance assessment, we propose in this paper a framework to evaluate the performance of EEG compression mechanisms in terms of post-compression seizure detection capability. In particular, we show that the above-mentioned 2-D algorithm can maintain diagnostic features down to bitrates of 2 bits per sample.