Fast wavelet transform for color image compression

In this paper a generalized algorithm for fast Harr wavelet transform was developed for the compression of monochrome and color images, and was evaluated through three quantitative measures: namely, mean square error, mean absolute error and signal-to-noise ratio. Excellent reconstructed images were obtained with high rates of compression. The computational advantage of this 2D FHWT over the conventional 2D FFT is 2.506 for an image of 1024/spl times/1024 pixels in size.