An Efficient Lossless Compression Algorithm for fMRI Data Volume

This paper presents an efficient lossless compression scheme designed for the fMRI data volume. An fMRI time series is composed of trend, noise and external stimulus response, which have different characteristics in wavelet domain. Reversible integer wavelet transform is applied to the fMRI time series. The smooth trend signals are extracted from the low frequency wavelet bands and they are coded by a lossy compression scheme. The lossy compression residual, noise, and other components are inversely transformed into time domain followed by adaptive clustering and then lossless coding. The experimental results show that the proposed method outperforms many existing schemes.