计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (21): 132-137.DOI: 10.3778/j.issn.1002-8331.1607-0226

• 模式识别与人工智能 • 上一篇    下一篇

改进的Gabor变换和二维NMF融合的人脸识别

王晓华,杨清梅,杨  涛   

  1. 西安工程大学 电子信息学院,西安 710048
  • 出版日期:2017-11-01 发布日期:2017-11-15

Face recognition based on improved Gabor transform and non-negative matrix factorization

WANG Xiaohua, YANG Qingmei, YANG Tao   

  1. School of Electronic Information, Xi’an Polytechnic University, Xi’an 710048, China
  • Online:2017-11-01 Published:2017-11-15

摘要: 为了得到高质量的人脸特征,提高人脸识别性能,提出基于改进的Gabor变换和(2D)2NMF(二维非负矩阵分解法)的人脸识别方法。改进的Gabor变换提取的特征有较高的品质,鲁棒性增强。二维非负矩阵分解法降维能大大降低图像数据维数,缩短计算时间,提高识别率。最后在ORL人脸库中进行实验,结果表明改进的Gabor变换和二维NMF方法相结合计算时间略微增加,但识别效率明显提高,从而证明了该方法的有效性。

关键词: 人脸识别, Gabor变换, 二维非负矩阵分解法

Abstract: In order to get the high quality facial features and improve the performance of face recognition, the face recognition method based on improved Gabor transform and two-dimensional non negative matrix factorization is proposed in this paper. Improved Gabor transform extracts the characteristics with a higher quality, enhances the robustness. Two dimensional non negative matrix decomposition of dimensionality reduction can greatly reduce the dimension of the image data, shorten the calculation time, improve the recognition rate. At last, experiments are carried in the ORL face database. The results show that improved Gabor transform and two-dimensional NMF method have a slight increase in the calculation time, but the recognition efficiency is improved obviously, thus proves the effectiveness of this method.

Key words: face recognition, Gabor transform, two-dimensional non-negative matrix decomposition method