【摘要】 脸部网格模型的建立是基于模型人脸合成技术中的关键步骤 .提出一个半交互式正面人脸模型调整算法 ,首先利用区域增长法和模板匹配自动确定图像中人脸及各特征区域的大致位置 ,然后交互地提取人脸特征点的准确二维位置 ,最后自动确定脸部姿态和调整模型非特征点 ,即获得输入人脸模型 .该算法简便实用 ,费时较少 ,具有实用价值更多还原

【Abstract】 Face modeling is an important step of model_based face reconstruction. We adopt a method to adapt the front facial wireframe model with some user’s interaction. At first we locate the positions of the face and the feature regions by region growing and template matching. Secondly we extract the exact positions of features by hand, and finally we calculate the rotation of the head and adapt the other nodes of the model using the extracted features and then to get the input face model. The test shows that this method is simple and useful.更多还原