Representation of Image Data by New Lifting based Biorthogonal Wavelets and Its Application to Lossless Compression

In this paper an attempt has been made to derive the lifting scheme for a set of new bi-orthogonal wavelets and apply on image compression. Four new bi-orthogonal wavelets are designed by taking different basis functions which are selected so as to capture the sharp edges which are common in images. For these classical wavelets, lifting versions are calculated and presented in this paper. The lifting step calculation is simplest among the schemes in related literature. From the designed wavelet filters, the poly-phase matrix is written and decomposed the matrix into a form from which lifting steps are directly available. To incorporate coding Set Partitioning In Hierarchical Trees (SPIHT) algorithm is used. The Peak Signal to Noise Ratio (PSNR), Compression Ratio (CR), Transforming Time (TT), Encoding Time (ET) and Decoding Times (DT) are calculated. It was found that the compression performance of new classical wavelets is a little better than that of existing classical/first generation wavelets and that of new lifting based wavelets is far better than that of existing classical wavelets and also existing lifting based wavelets.