الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 97 |  |  |
| | | from | 97 | | |
Abstract
3D Mesh compression algorithms mostly are driven by the mesh connectivity code. The geometry representation is usually employs a straightforward prediction method applied to the vertex sequence as dictated by this connectivity code. This generates a suboptimal geometry code, which results in significant loss in code efficiency, since the geometry dominates the mesh information content. Many redundancies are observed in this code. Such as, each vertex could be repeated in worst case 6 times (internal vertex). This work proposes a structure approach to arrange the vertices of a mesh in a label manner,
Moving between meshes, in other words, from one mesh to another requires direction and vertices. from geometric point of view the best mesh is a triangle shape. The triangulation selection based upon quantization level. In this work preprocessing is the first phase, in this phase the input is 3D solid object represented as a huge number of vertices. Normally to represent this object we need to call these vertices with its polygons which take a large storage capacity. The output from this phase is a lookup table for the vertices, each one with a label number, and the three coordinates (x, y, and z). The meshes are represented by the best optimal triangles. The second phase is the main compression phase.
The main objective of this thesis is to develop powerful new compression technologies to reduce storage requirements and transmission times from static and dynamic 3D models.
A modified edge breaker algorithm is implemented using matlab and c++. The output of this phase is the same vertices and ASCII file which has a sequence of direction (CLERS). In other words, filling the area with a group of triangle (the best define face) by moving from face to another using a lookup table which build on five directions (CLERS). The input of this phase is the OV table file. The inverse procedure is introduced, in other words, in the receiving channel. This part converts the files CLERS and OV table to the original 3D Model. The reconstructed model is compared with the original 3D Model.
Our proposed algorithm gives the best result, in other words the minimum percentage error (0.03446). The reason behind this is the modification of the edge breaker which results in almost no errors. The modification is done mainly by enhancing the role of the preprocessing phase.