Unified Image Compression Using Reversible and Fast Biorthogonal Wavelet Transform

Abstract

Wavelet transforms are known to have excellent energy compaction characteristics and are therefore ideal for signal and image compression. Although this approach has been vigorously developed during the recent years, general orthogonal wavelet transform filters and most of sub-band coding filters have many taps and require many floating point multiplications. Fast image compressors are presented using biorthogonal wavelet transforms, which gives high computational speed and excellent compression performance. Special spline biorthogonal wavelets are used where the filter coefficients are dyadic rational numbers, and convolutions with these filters can be performed by using only integer arithmetic shifting and addition operations. Following the transform, the Hilbert scanning is used for encoding to gain additional compression.