Wavelet transforms plays an important role in image compression techniques that developed recently. Here three 2-D functions are considered which are approximated and compressed using multilevel discrete 2-D wavelet transforms like Haar, Daubechies, Coiflet and Symlet. The images that are compressed are tested for quality using the error metrics like mean square error (MSE), peak to signal noise ratio (PSNR), maximum error (MAXERR), L2RAT, compression ratio and bit per pixel. The evaluation of the above mentioned wavelets is synthesized in terms of experimental results which demonstrates that Haar wavelets provides high compression ratios for 2-D exponential functions and the product of sine and cosine functions whereas Daubechies wavelet gives good compression ratio for 2-D periodic function.
Read full abstract