针对光照对人脸识别影响的问题,提出一种结合小波变换和光照补偿的人脸识别方法.该方法首先利用离散小波变换将人脸图像的低频子带和高频子带分离,在小波变换的低频子带上分别进行直方图均衡化和对数变换,将处理后的低频子带进行融合构成新的低频子带.接着对高频子带进行阈值去噪后乘以一个标量,构成新的高频子带.最后利用小波逆变换重构出新的人脸图像并利用PCA算法进行识别.实验结果表明,该方法能有效地削弱光照的影响,提高人脸识别率.%Aiming at the effect of illumination on face recognition, a new face recognition method combining wavelet transform and illumination compensation is proposed.Firstly, the low-frequency sub-band and high-frequency sub-band are separated by discrete wavelet transform, then the histogram equalization and logarithm transformation are carried out on the low-frequency sub-band of the wavelet transform, and the processed low-frequency sub-band is merged to constitute a new low-frequency sub-band.Then, the high-frequency sub-band is denoised by threshold and then multiplied by a scalar to constitute a new high-frequency sub-band.Finally, a new face image is reconstructed using the inverse wavelet transform and the PCA algorithm is used to identify the face image.Experimental results show that this method can effectively reduce the influence of illumination and improve the recognition rate of face.
展开▼