Wavelet-based progressive fast recompression technique for 3D large mesh data

Abstract

3D mesh is one of the representation of 3D object defined as a combination of geometric primitives, utilized in 3D printing, modeling, and reconstructing 3D objects by scanning. Recently, increasing interest has appeared on handling mesh data in network environments such as Web3D and in this case multi-resolution scheme can be useful for data transmission, which has been studied in terms of progressive compression. For large models, it is effective for sending quick contour view first but the compression time becomes much higher. Whenever we needs deformation, recompression takes as much time as the previous time elapsed. In this thesis, we greatly reduced the recompression time when connectivity-preserving deformation happened. We first compressed a large mesh progressively by wavelet-based method, and then extracted and directly modified wavelet coefficients affected by deformed area using our propagation algorithm. Our method increased the recompression speed more than 15 times to the original compression speed, even if the entire mesh was deformed. We also showed that the speedup increased up to 33 times when local deformation acted on local region of lower than 5% of mesh.