首页> 外国专利> DETECTION PROCEDURE OF CHARACTERISTICAL FACE POINTS ON THE SEQUENTIAL SEARCH BASIS WITH DIFFERENT REGRESSION LINEAR MODELS

DETECTION PROCEDURE OF CHARACTERISTICAL FACE POINTS ON THE SEQUENTIAL SEARCH BASIS WITH DIFFERENT REGRESSION LINEAR MODELS

机译:不同回归线性模型的序列搜索基础上特征面点的检测过程

摘要

a method for detection of characteristic points of a face in a digital image, to determine the sought after facial points using serial search with the crowd in front of the learned linear regression models.the procedure is constructed and the teaching work, on the basis of the learning data set of mixtures of linear regression models, and the use phase, in which the trained regression models applied to mixtures obdelovani figure to determine the location of characteristic points of the face.search for the characteristic points is divided into successive stages so that an initial assessment of the position of the characteristic point set with the result of detection of faces, based on the regression in each step of the update, and gradually improves.it used a mixture of regression models on the basis of the pripadnostne functions in each step of the sequence to karakteristikam current images and thereby provides effective and robust detektiranje facial characteristic points in the image.
机译:一种用于检测数字图像中人脸特征点的方法,通过在学习的线性回归模型之前使用人群进行串行搜索来确定所需的面部点。在此基础上,构造了程序并进行了教学工作线性回归模型的混合学习数据集和使用阶段,其中训练有素的回归模型应用于混合obdelovani图以确定人脸特征点的位置。将特征点搜索分为连续阶段,以便根据更新的每个步骤中的回归对特征点集的位置进行初步评估(基于面部检测的结果),并逐步进行改进。在每个过程中使用pripadnostn​​e函数基于混合回归模型序列中的步骤,以生成当前图像,从而在图像中提供有效且强大的detektiranje面部特征点。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号