Research Of Facial Data Normalization Based On Craniofacial Partition

Nowadays, craniofacial restoration and statistical analysis are widely used in craniofacial morphological.These research and application require skull and 3D mesh models having better quality and higher quantity. However, in the relevant research,3D craniofacial mesh models have much difference in number of point, topology structure and memory size. It can't meet the needs of craniofacial morphology statistical analysis. To solve this problem, this paper carried out a technical study on craniofacial normalized based on data partitions, mainly characterized by craniofacial template point calibration, zoning and simplify operation, get the standard craniofacial model, then use corresponding method based on geodesic, in order to implement registration between new model and standard model. At last, all the craniofacial data should have the same data points and structure. The main research work includes:1. Use geometry as a template to achieve craniofacial model data partition. Partition the craniofacial uses template close to human facial. The experimental results show that this method has high efficiency, the boundary is smooth and the partition result is more accurate.2. Simplify the 3D model by a craniofacial data method based on shape analysis. After the facial feature analysis, some key points outside of the feature points are retained in the process of simplification. Different regions should be taken in different simplification criteria. In a limited number of points, it can describe the structure and characteristics more completely.3. Propose the craniofacial corresponding method based on geodesic. Establishing corresp-ondence between new craniofacial and standard craniofacial by expanding the initial point set.Using this approach can solve the problem of speed slowly and position isn't unified. What's more, the method in this paper has a higher accuracy.4. Design and develop a three-dimensional craniofacial data normalized prototype system. This system is easy to use, with good interaction, and it can process new craniofacial data rapidly. The models after processing have same number of points and topology. Craniofacial data normalization provides a reliable foundation for the further research.