With the growing demand for image processing, efficient image data compression techniques are much needed for image data transmission and storage. Two image data compression techniques based on the Laplacian Pyramids are examined: the Laplacian Pyramid Vector Quantization (LPVQ) and the Laplacian Pyramid Predictive Compression (m). The LPVQ can provide a reduction from 8 bits per pixel (bpp) to 0.45 bpp with negI.igi b1.e degradation in image quality ( 0.6% rms reconstruction error), as shown in a test, j-mage. The LPFC can provide comprable performance with a proper choice of image model. Both techniques thus can be superior to existing transform ccding and predictive compression techniques. ‘I‘he problems of additive noise, pre-and post-processing and hardware realization are also considered.

http://ieeexplore.ieee.org/iel6/8363/26345/01169928.pdf

N. Dartmouth

Papers

LAPLACIAN PYRAMID IMAGE DATA COMPRESSION C.H. Chen