Determining suitable wavelet filters for visual sensor networks

Abstract

Visual sensor networks (VSN) require strict constraints on energy consumption. So, low power wavelet-based coder (WBC) is becoming crucial in VSN design. This makes more difficult the determination of appropriate wavelets. An appropriate wavelet is a wavelet that consumes few energy during image processing while permitting an acceptable quality of the reconstructed image at the reception. In this paper, we make a comparative study between DWT transforms because DWT has several advantages. We consider two known ways of implementing DWT: the classical convolutional-based wavelets and the relatively new lifting-based wavelets.