SALIENCY GUIDED WAVELET COMPRESSION FOR LOW BITRATE IMAGE AND VIDEO CODING

Abstract

We propose an elevated saliency guided wavelet compression scheme for low-bitrate image/video coding purposes. Primary areas (faces in protection digicam feeds, automobiles in site visitors surveillance) get degraded greatly at low bitrates with the aid of existing compression specifications, similar to JPEG/JPEG2000/MPEG-four, when you consider that these do not explicitly utilize any expertise of which areas are salient. We design a compression algorithm which, given an image/video and a saliency value for every pixel, computes a corresponding saliency worth within the wavelet turn out to be domain. Our algorithm ensures wavelet coefficients representing salient regions have a high saliency price. The coefficients are transmitted in reducing order of their saliency.This makes it possible for primary regions within the photo/video to have excessive constancy even at very low bitrates. Additional, our compression scheme can handle a couple of salient areas with exceptional relative value. We evaluate the performance of our approach with the JPEG/JPEG-2000 image necessities and the MPEG-4 video general by means of two experiments: face detection and automobile monitoring. We show extended detection premiums and pleasant of reconstructed photographs/videos making use of our Saliency established Compression (SBC) algorithm.