Study on wavelet spectral band based EEG compression

The functioning of the brain is mainly controlled by nervous system. It sends out stimulus in the form of electrical pulses to different parts of the brain. These electrical pulses are recorded by a procedure known as Electroencephlography. EEG signal acquisition usually last for 2-8 hours, or can be extended over a week. This places a bigger problem on storage requirements as well as transmission of EEG. So, EEG signal compression is attaining more interest on medical field. EEG data needs to have more clinical importance as it deals with a patients health, hence requires an effective compression system involving perfect reconstruction of signal. The proposed method includes processing, band separation of EEG signal and identifying wavelets suitable for efficient compression of EEG. EEG signal is processed using various wavelet families like symlet, coiflet, biorthogonal and daubechies wavelets. For integer order compression, db5 wavelet produces much effective compression than any other wavelets.