Analysis of joint predictive-transform coding for still image compression

  • Authors:
  • Lorenzo Cappellari;Gian Antonio Mian

  • Affiliations:
  • University of Padova, Department of Information Engineering, via Gradenigo 6/B, 35131 Padova, Italy;University of Padova, Department of Information Engineering, via Gradenigo 6/B, 35131 Padova, Italy

  • Venue:
  • Signal Processing - Signal processing in communications
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents a performance analysis of hybrid predictive-transform coding for compression of both one-and two-dimensional signals. Our aim is to investigate whether recently developed H.264/AVC intraframe coding paradigm outperforms previous DCT based video compression schemes. An increased coding gain should be expected, but, while for one-dimensional signals the improved coding gain offered by this hybrid approach is easily shown, extension to two-dimensional cases is not straightforward. In this case, however, higher performance is still justified using a new model for local correlation. To provide evidence for its application, we compared the proposed parametric model with measured correlation data and proved that it approximates real data better than isotropic or separable models do.