Adaptivity with near-orthogonality constraint for high compression rates in lifting scheme framework

摘要: Since few years, Lifting Scheme has proven its utility in compression field. It permits to easily create fast, reversible, separable or no, not necessarily linear, multiresolution analysis for sound, image, video even 3D graphics. An interesting feature of lifting scheme is the ability build adaptive transforms compression, more than with other decompositions. Many works have already be done this subject, especially lossless near-lossless compression framework : better usually used methods can obtained. However, most techniques near-lossless extended higher lossy rates, simplest cases. Indeed, due quantization error introduced before coding, which controlled propagation through inverse transform. Authors put their interest classical Scheme, linear convolution filters, but they studied criterions maintain a high level adaptivity and good This article aims present relatively simple criterion obtain filters able image rate, tested here Spiht coder. For this, upgrade predict are simultaneously adapted thanks constrained least-square method. The constraint consists near-orthogonality inequality, letting sufficiently adaptivity. Some results given, illustrating relevance method, short filters.