计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (22): 150-155.DOI: 10.3778/j.issn.1002-8331.1707-0505

• 模式识别与人工智能 • 上一篇    下一篇

面向人脸识别的WPD-HOG金字塔特征提取方法

刘文培1,李凤莲1,张雪英1,田玉楚1,2   

  1. 1.太原理工大学 信息工程学院,山西 晋中 030600
    2.昆士兰科技大学 电机工程及计算机科学学院,澳大利亚 昆士兰
  • 出版日期:2018-11-15 发布日期:2018-11-13

WPD-HOG pyramid feature extraction method for face recognition

LIU Wenpei1, LI Fenglian1, ZHANG Xueying1, TIAN Yuchu1,2   

  1. 1.College of Information Engineering, Taiyuan University of Technology, Jinzhong, Shanxi 030600, China
    2.School of Electrical Engineering and Computer Science, Queensland University of Technology, Queensland, Australia
  • Online:2018-11-15 Published:2018-11-13

摘要: 人脸识别技术可应用于各监控和安保领域,它涉及特征提取、识别模型等关键技术。其中特征提取方法直接影响识别效果,目前所用的特征提取方法存在特征表达不全面、计算复杂度高等问题。据此,提出一种基于WPD-HOG金字塔的人脸特征提取方法,该方法结合小波包分解(Wavelet Packet Decomposition,WPD)、图像金字塔以及方向梯度直方图(Histograms of Oriented Gradients,HOG)对人脸图像特征进行有效表征,最终将WPD-HOG金字塔特征通过SVM分类器进行分类。通过在ORL人脸库上进行实验,与四种对比方法HOG、HOG金字塔、FWPD-HOG以及FWPD-HOG金字塔进行比较,实验结果表明,WPD-HOG金字塔特征提取方法的识别率要高于对比方法,且在噪声方面具有较好的鲁棒性。

关键词: 人脸识别特征提取, 小波包分解, 图像金字塔, 方向梯度直方图

Abstract: Face recognition technology can be applied in the field of monitoring and security, which involves key technologies such as feature extraction and recognition model. The feature extraction method has a direct influence on the recognition effect. At present, the feature extraction method has the problems of incomplete expression and high computational complexity. For solving this problem, this paper proposes a kind of facial feature extraction method:WPD-HOG pyramid. The WPD-HOG pyramid feature extraction method combines the Wavelet Packet Decomposition(WPD), image pyramid and Histograms of Oriented Gradients(HOG) together to characterize the face image feature. Finally, the WPD-HOG pyramid features are identified by the SVM classifier for face recognition. Experiments are conducted over ORL data set to demonstrate the proposed approach. Compared with the four benchmark methods:HOG, HOG pyramid, FWPD-HOG and FWPD-HOG pyramid, the experimental results show that the recognition performance, computation complexity and noise robustness of the proposed method are the best.

Key words: face recognition feature extraction, Wavelet Packet Decomposition(WPD), image pyramid, Histograms of Oriented Gradients(HOG)