【作者】 韩璐；

【导师】 段哲民；

【作者基本信息】 西北工业大学 ， 电路与系统， 2007， 硕士

【摘要】 面部轮廓整形图像的参数测量与分析对面部轮廓整形手术的成功率以及术前准备、术后效果评估均有重要作用。以往的面部参数测量大多基于造价昂贵的医学成像设备，如计算机断层成像设备(Computed tomography，CT)、核磁共振设备(Magnetic resonance imaging，MRI)等，而且专业的面部轮廓测量软件也大都基于这些设备得到的面部图像。由于基于这些设备获得的图像成本较高，使得这些专业软件难以推广。 鉴于此，本论文以普通数码相机获取的面部图像为素材，首先对图像进行预处理，增强面部轮廓信息，同时去除不需要的信息；其次利用基于遗传算法的特定轮廓分割方法对预处理得到的图像进行轮廓分割，得到了精确的二维面部轮廓模型；接着运用三维重建的方法对得到的二维轮廓模型进行三维重建，得到面部的三维轮廓模型；最后以二维和三维模型为基础，运用聚类分析等数学方法对轮廓参数进行分析和测量，得到了面部整形手术中最重要的两项参数，空间下颌角度与升支／下颌比。 经过实测，得到的结果准确性良好。本论文提供了一种基于普通数码图像的面部参数测量方法，实验结果证明该方法鲁棒、有效。更多还原

【Abstract】 The analysis and meterage of mandible parameter play an important role in facial plastic surgery. The expensive equipments, such as Computed tomography (CT), Magnetic resonance imaging (MRI), etc, are employed to get the face structure images, which cost a lot in whole process. Furthermore most professional software for the surgery is almost based on the images caught by those expensive machines and it is hard to be popular due to the high cost of the images.This paper present a novel way to complete the analysis and meterage based on the digital images caught by normal digital camera. First, preprocessing, such as image enhancement, edge detection, is applied to make images adapt to the mandible contour segmentation. Second, the given contour segmentation algorithm based on genetic algorithm is put forward to extract the 2D mandible contour accurately. Third, the 3D contour model is reconstructed by computer vision technology. Last, the two most important parameters of mandible are computed: the dimensional mandibular angle and ratio between mandibular ramus and mandible.By means of practical measure, the computed results are proved correct. This paper provides a new mandible parameters measurement method which is based upon normal digital images. The experimental results prove that it is robust and efficient way.更多还原