高级检索
税午阳, 周明全, 武仲科, 邓擎琼. 数据配准的颅骨面貌复原方法[J]. 计算机辅助设计与图形学学报, 2011, 23(4): 607-614.
引用本文: 税午阳, 周明全, 武仲科, 邓擎琼. 数据配准的颅骨面貌复原方法[J]. 计算机辅助设计与图形学学报, 2011, 23(4): 607-614.
Shui Wuyang, Zhou Mingquan, Wu Zhongke, Deng Qingqiong. An Approach of Craniofacial Reconstruction Based on Registration[J]. Journal of Computer-Aided Design & Computer Graphics, 2011, 23(4): 607-614.
Citation: Shui Wuyang, Zhou Mingquan, Wu Zhongke, Deng Qingqiong. An Approach of Craniofacial Reconstruction Based on Registration[J]. Journal of Computer-Aided Design & Computer Graphics, 2011, 23(4): 607-614.

数据配准的颅骨面貌复原方法

An Approach of Craniofacial Reconstruction Based on Registration

  • 摘要: 颅骨面貌复原技术在法医刑侦领域具有重要应用,为了估计给定颅骨的未知面貌,提出点云配准与五官(鼻、眼睛、嘴)配准结合的颅骨面貌复原框架及基于软组织厚度的评价方法.首先构建颅面数据库并进行坐标校正;然后确定配准点集及其对应关系,采用ICP算法和薄板样条函数两次配准实现参考颅骨与待复原颅骨的精确配准,并将参考模型的软组织厚度作为待复原颅骨的软组织厚度实现面貌初步复原;最后提出一种五官配准方法以提高面貌复原的准确性.提出基于软组织厚度比较的评价方法来验证面貌复原结果的正确性.实验结果证明,文中方法可以有效地实现给定颅骨的面貌估计.

     

    Abstract: Craniofacial reconstruction has important applications in the domain of forensic science.In order to estimate the appearance of the unidentified skull,this paper proposes a novel framework of craniofacial reconstruction by combining dense point sets registration and organs(nose,eyes and mouth) registration.We introduce an evaluation method of craniofacial reconstruction based on the soft-tissue thickness.Firstly,a 3D model database of faces and skulls is built and processed.Secondly,registration point sets are selected and their correspondence is established.The iterative closest point(ICP) and thin-plate spline(TPS) are adopted to achieve registration from the reference skull to the unidentified skull.Furthermore,the soft-tissue thickness of the reference skull is taken as the initial craniofacial estimation of the unidentified skull.Finally,a method of organ registration is presented to improve accuracy of the craniofacial reconstruction.In order to verify the correctness of the algorithm,the soft-tissue thickness is compared between the final result and the reference.The feasibility and usability of the algorithm are confirmed through our experiments.

     

/

返回文章
返回