Contour Extraction And Contour-based Craniofacial Model Reconstruction Algorithm Research

In the craniofacial reconstruction process, a large number of craniofacial models are mostly generated by MC method, but the models generated using this algorithm takes too much space for storing and most of them has indelible impurities which cause the problem of insufficient accuracy when simplificated to a certain degree. In order to solve this problem, This paper tries to reconstruct better craniofacial models based on the key problem about contour extracting technology. The research in this paper is mainly about:1. Extract craniofacial contour. proposed a method which merge Snake algorithm and ray method to accurate extract the complex, discontinuous outermost layer of the skull contour.2. Contour processing. achieved the point sort algorithm, to make sure the contour points is stored according to an order. then we can do the three-dimensional model reconstruction from planar contour smoothly.3. Craniofacial surface reconstruction. proposed an improved method of the shortest diagonal method to connect closed face contour lines and discontinuous skull contour lines between layers, reconstruct no thickness and impurity craniofacial models. Result shows it provides excellent data support for craniofacial reversion.4. Model optimization. Design a "paint brush" to remove impurities, while use point deletion algorithm to simplify triangle mesh and use the linear interpolation to repair grid hole.Develops a CT image processing and three-dimensional reconstruction system. In the contour extraction stage, the proposed algorithm has the advantage of choosing which part of the contours been extracted and only extracting the outermost contour compared with the traditional algorithm. In the reconstruction stage, The proposed algorithm can be smart on connecting contour lines. The experimental results show that the algorithm in this paper can generate models which occupy a small hard disk space and express a lot of information. It also provides an excellent platform for further craniofacial morphology study.