MULTIRESOLUTION DECOMPOSITION AND RECONSTRUCTION OF COLOR IMAGES

Abstract

The multiresolution nature of Discrete Wavelet Transform is a proven powerful tool to represent color images, decomposed along the horizontal and vertical directions. These subimages correspond. , to different frequency hands. This technique is intrinsically superior to other transform techniques as it takes into account psychovisual features both in the space and frequency domain. In this dissertation, several 512 x 512 color images were taken and the Discrete Wavelet Transform is applied to obtain the subimages at different resolutions. Inverse Wavelet Transform is used to reconstruct the image from the subimages. Software implementation of Wavelet Transform has been described. It has been observed that for all the images the PSNR value is least for the red component and maximum for the blue component respectively. Findings suggest that the quality of the reconstructed images are fairly good in terms of the information content.