Author(s)

Jianlin Chen, Gaoyong Luo, Fasheng Zhou, Haitao Cao

Corresponding Author

Gaoyong Luo

Abstract

Image denoising has played an important role in face recognition. Face features are more associated with edge detection. However, noise reduction methods such as adaptive Wiener or Kalman filtering based on minimizing mean square error achieve suboptimal results but at the same time blur image by filtering out high frequency contents related to edge. This paper presents a spatial-frequency image filtering method by wavelet decomposition to achieve better edge preservation while reducing noise significantly. Experimental results show that the proposed method performs better than other adaptive filtering methods for feature extraction of face recognition by neural network with multilayer perceptron.

Keywords

Face Recognition, Wiener Filter, Kalman Filter, Wavelet Filter, Neural Network.