The Progressive 3D Geometry Compression Technologies Based-on Degree Driven

Recently, the three dimensional geometry data are required to inspect and manipulate in real time at local/remote place for many applications, in order to enhance the human productivity and satisfaction. On the other hand, the number and the complexity of the three dimensional geometry data sets are growing rapidly, to meet the need of accurate precision and complicated details. Consequently, a new geometry coding scheme, 3D mesh compression, is proposed to solve the bottlenecks in the mesh storage, process, transmission and render. 3D mesh compression, first proposed by Deering, is a compact data representation with the following properties, such as fewer data quantity with compressed format, faster rendering speed, and better applicability to the transmission. To deal with the difficulties of large data storage and scheduler between main memory and disk, the former works concentrate on the single-rate approach, which treats the attributes (e.g. connectivity and geometry) as a whole. While the recent researches mainly simplify the input mesh into levels of detail and progressively transmit them. The advantage of progressive compression is that user can achieve an approximate mesh even if a few bits have been received.Firstly, all algorithms, which compress 3D meshes efficiently, have been examined and classified into single-rate compression and progressive transmission type. In brief, main ideas behind each class were described, and the advantages and shortcomings of the algorithms in each class were compared.Secondly, a new progressive compression for lossless transmission of triangle meshes is proposed, which is improved from the progressive valence-driven compression. Based on the facts that a mesh is composed of connectivity, geometry, and parameter information, a new method of multi-granularity quantization in geometry encoding is presented, which allows making better of the dynamic range (different number of bits in normal/tangential components) than with a fixed-bit representation. Consequently, with the aid of the anchors and variational geometry coding, the rate-distortion performance is much better than the original and the recent similar lossless algorithms, even can compare beauty with spectral compression techniques at low bit rates. Besides these, the experiments have also thrown that the proposed algorithm can achieve a better coding efficiency..Thirdly, the distributed communication framework has been examined, which provides the foundation for the 3d mesh compression algorithm on the current network. A new distributed framework has been proposed, whose client is designed by the model-view-controller design pattern, and whose server is service-oriented type. The framework has many advantages, such as short response time for the remote user, more vivid visualization, and easier expandability. At last, a conclusion is made, and some research points are foreseen.