This paper presents a novel approach to extract image and edge details that
are useful for segmentation of video objects in object-based coding. The p
roposed approach provides a transform domain technique and uses perceptual
properties and edge contrast information in the transform coefficients alon
g with the human visual system weighting to extract image and edge details
that an perceptually most significant. (C) 2000 Elsevier Science B.V. All r
ights reserved.