Sparsity-based optimization of two lifting-based wavelet transforms for semi-regular mesh compression

This paper describes how to optimize two popular wavelet transforms for semi-regular meshes, using a lifting scheme. The objective is to adapt multiresolution analysis to the input mesh to improve its subsequent coding. Considering either the Butterfly- or the Loop-based lifting schemes, our algorithm finds at each resolution level an optimal prediction operator P   such that it minimizes the L1-normL1-norm of the wavelet coefficients. The update operator U is then recomputed in order to take into account the modifications to P. Experimental results show that our algorithm improves on state-of-the-art wavelet coders.

Graphical abstractFigure optionsDownload full-size imageDownload high-quality image (87 K)Download as PowerPoint slideHighlights► We explain how to optimize two wavelet transforms for semi-regular meshes, using a lifting scheme. ► Our algorithm optimizes the prediction of the Butterfly-based scheme according to the input mesh. ► We explain how to recompute the update operator in function of the optimized prediction operator. ► We then expand this algorithm to a Loop-based lifting scheme. ► We show that our algorithm improves the wavelet coding of specific meshes.