基于正方形组合描述符与深度森林的三维人脸分割方法  

3D face segmentation method based on square combination descriptor and deep forest

作　　者：杜星悦 董洪伟[1] 杨振 DU Xing-yue, DONG Hong-wei, YANG Zhen(College of The Internet of Things Engineering, Jiangnan University, Wuxi 214122, China)

机构地区：[1]江南大学物联网工程学院,江苏无锡214122

出　　处：《传感器与微系统》2018年第9期35-37,共3页Transducer and Microsystem Technologies

摘　　要：三维人脸分割技术在三维人脸研究领域有实际意义,其结果对三维人脸的特征点定位,表情识别,人脸动画等领域作用较大。针对目前已有的分割算法在精度上存在一定不足,提出了基于正方形组合描述符以及深度森林的三维人脸分割方法。对三维人脸进行分割实验,结果表明:所提算法精度高,分割效果良好,具有实际应用意义。The three dimensional（ 3 D） face segmentation technology has practical significance in the field of 3 D face research. The result is important for the 3 D face feature points localization,expression recognition,facial animation,etc. Because of the shortage in precision of the existing segmentation algorithm,3 D face segmentation method based on square combination descriptor and deep forest is proposed,and segmentation experiments are carried out. The experimental results show that this algorithm is of high precision,and good segmentation results,which has practical application significance.

关 键 词：三维人脸 三维分割 正方形组合描述符 深度森林 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]