Wavelet coding of 3-D shape data using space-frequency quantization

Summary form only given. Efficient representations of three dimensional object shape have attracted wide attention for transmission of computer graphics data and for interactive design in manufacturing. Polygonal meshes which consist of connectivity data of the vertices and their geometry data are often used for representing an object shape. The connectivity of the vertices is efficiently coded with generalized triangle stripes and several generalized meshes. However the geometry data with a larger value than the connectivity data is compressed by changing the step size of scalar quantization and applying one-dimensional prediction coding. We propose a data compression method applying the wavelet transform and hierarchical quantization for geometry data structured by a quadrilateral mesh. The geometry data is transformed to multi-scale wavelet coefficients of the same number as original data by applying a shape-adaptive wavelet transform along two axes of the quadrilateral mesh. The wavelet coefficients are entropy coded by optimal combined zerotree quantization, which takes into account dependency of the coefficients at similar position in each scale, with scalar quantization of the remaining coefficients. The performance of the proposed method is evaluated using two metrics based on shape distortion and normal vector distortion of reconstructed geometry data. The experimental results using range data of a statue surface measured at the same intervals along circular and altitude directions in circular cylindrical coordinates show that the new method has good performance compared to conventional methods