Image compression using biorthogonal wavelet transforms with multiplierless 2-D filter mask operation

We present image compression algorithms using biorthogonal wavelet transforms with dyadic regional filter coefficients, which provide high computational speed and excellent compression performance. Two-dimensional reconstruction filter masks are constructed to minimize data access operations and process only the nonzero-valued pixels when performing the inverse wavelet transform. Convolutions with these masks can be performed by using only arithmetic bit-shifting and integer addition operations. Additional compression gain is achieved by using Hilbert scanning for the preprocessing of encoding process.