一种复杂背景下的人脸特征提取算法

通过人脸特征提取是计算机人脸识别技术中的核心技术。主动轮廓模型方法(Activecontourmodel,也称Snakes方法)被广泛应用于特征提取、图像分割等技术中。常规Snakes方法应用于人脸特征提取鲁棒性不强,并且精度不高。对常规Snakes方法进行了改进,其改进主要包括两个方面:1)针对人脸肤色的特点改造了Snakes能量函数的图像能量项,用以提高算法的精度;2)基于人脸五官形状特点,增加了一个新的能量项———形状限定能量项,以期获得更好的鲁棒性。最后,对改进前后的Snakes方法进行了实验仿真与比较、对比研究。实验结果表明,改进后的Snakes方法具有良好的收敛性,并能准确收敛到人脸特征处。

An Algorithm for Facial Features Extraction in Complex Background Image

Facial features xetraction have become a kdy in the fislds such as face recognition. Active contour model (usually called as snakes method) has been widely used in the technology of image segmentation, feature extraction et al. In this paper, we improved the Snakes method mainly in two aspects. Firstly, for a better precision, we reconstructed the image energy item in the Snakes' energy function according to the complexion characteristics of human face. Secondly, a new energy item, the shape restriction energy item, is added to the Snakes function to restrict the shape of Snakes during the Snakes' iteration computation, to prevent the Snakes curve from converging mistakenly.Finally,to verify the validity of our improvements,we developed some programs based on our algorithm. Results from our experiments show that the Snakes curve produced by our improved Snakes method can fit the shapes of the facial features better and more quickly.