Comparison of Lifting and B-spline DWT Implementations for Implantable Neuroprosthetics

The discrete wavelet transform (DWT) has been shown to provide exceptionally efficient data compression for neural records. This paper describes area-power minimized hardware implementations of the lifting and B-spline DWT schemes for multi-level, multi-channel DWT and analyzes their performance tradeoffs for implantable neuroprosthetics. The lifting scheme is shown to be increasingly superior for a larger number of input channels.