基于活动轮廓模型的人脸特征信息标定方法

Facial Feature Information Calibration Method Based on Active Contour Model

针对传统人脸特征信息标定方法存在标定误差大的问题,笔者提出基于活动轮廓模型的人脸特征信息标定方法,首先根据二维空间曲线的演化原理创建活动轮廓模型,然后收集待标定的人脸图像样本,并通过去噪、增强、二值化处理实现人脸图像预处理,最后确定人脸轮廓的初始位置,定位提取人脸特征信息,实现对人脸特征信息的标定。实验结果表明,基于活动轮廓模型的人脸特征信息标定方法的误差更低。

Aiming at the problem of large calibration errors in traditional face feature information calibration methods,the author proposes a face feature information calibration method based on active contour models.First,an active contour model is created based on the evolution principle of two-dimensional space curves,and then the faces to be calibrated are collected.The image samples are preprocessed by denoising,enhancement,and binarization.Finally,the initial position of the face contour is determined,and the facial feature information is located and extracted to achieve the calibration of the facial feature information.The experimental results show that the calibration method of facial feature information based on active contour model has lower error.