Research On Spatio-Temporal Segmentation Based 3D Mesh Sequences Compression And Optimization

With the recent advances of animation production technology and 3D presentation technology,effective compression of 3D mesh animation data has been increasingly used in a variety of multimedia systems including virtual reality,gaming,etc.,especially in which requires remote transmission,display and storage of animation data.The compression of various dynamic mesh sequence data also has become an important topic in computer graphics community.This paper first presents a new spatiotemporal segmentation-based approach for the adaptive compression of the dynamic mesh sequences.Given an input 3D mesh animation,first utilizing the Kernel Canonical Correlation Analysis(k CCA)based Maximum Mean Difference(MMD)algorithm,compute an initial temporal cut to obtain a small subsequence by detecting the temporal boundary of dynamic behavior.Then,applying a two-stage vertex clustering on the resulting subsequence to classify the vertices into groups with optimal intra-affinities.After that,we design a temporal segmentation step based on the variations of the principle components within each vertex group prior to performing a PCA-based compression.Finally,applying a lossless compression of the PCA bases and coefficients for complete features retaining.The approach can adaptively determine the temporal and spatial segmentation boundaries in order to exploit both temporal and spatial redundancies.This paper also proposes a boundary edition method and a spectral clustering-based dynamic reshaping model that is performed on spatio-temporal segments to enhance the compression of 3D mesh sequences.The boundary editing step modestly expands the space-time segmentation block,the segmentation boundary is stored in two adjacent data matrices at the same time,ensuring the boundary consistency during decompression.For matrix reshaping,after the lossy compression of spatio-temporal segments through Principal Component Analysis(PCA),the spectral clustering of all PCA decomposition elements is calculated.Then,three novel reshaping schemes(namely,Row-wise matrix scheme,Arch-wise matrix scheme,and Curl-wise matrix scheme)of the PCA elements within each cluster are proposed.Through extensive experiments and comparisons,the results show the two kind of optimization model can substantially improve the compression performances on various 3D mesh sequences.In a word,this paper develops an adaptive spatio-temporal segmentation approach for 3D mesh animations by exploiting the spatial and the temporal redundancies simultaneously based on the dynamic behaviors,the adaptability enhances the robustness of the algorithm when facing various data.Enhance methods based on boundary edition and matrix reshaping are proposed,through quantitative comparison and visualization of reconstruction errors,experiments show that the space-time segmentation method and its optimization algorithms are effective and competitive for the compression of 3D mesh sequences.