An efficient image compression method based on pyramid sub-band filter
ing is presented. Gaussian and Laplacian pyramids are adopted to const
ruct a sub-band filter. The filter is then used to decompose an origin
al image signal into three sub-bands: low-frequency, middle frequency
and high frequency. The sub-bands are separately encoded by different
schemes according to the statistical characteristics of the sub-bands
and the properties of the human visual system. Simulation results show
that the compression efficiency is improved significantly.