Quadrature-based structure with refined context-based entropy coding for subband image compression

In this paper, we propose a simple subband image compression scheme using wavelet decomposition. The proposed coding scheme employs a multi-level dyadic wavelet decomposition, organizes and orders the resultant wavelet coefficients using pre-determined quadtree data structure, and finally encode symbols by arithmetic coding with pre-designed context information. Despite the simplicity of the scheme, the experimental result shows it is competitive with the presently available schemes when compared with various types of test images.