基于小波变换人脸图像融合方法的研究

Research of face image transform based on wavelet

利用小波高频系数区域选大法、小波区域能量选大法、小波系数区域标准差选大法、小波区域平均梯度选大法四种融合算法,通过小波变换将图像金字塔选择系数进行一系列的处理后构成融合图像信息。最后从均值、信息熵、交叉熵、扭曲程度、相关系数等方面对融合后图像进行了评价,验证了算法的可行性。

Using four fusion algorithm such as high frequency wavelet coefficients of regional election,selection of wavelet regional energy,standard deviation of wavelet coefficients,and selection of wavelet regional average gradient,the image information is formed by means of wavelet transform.Finally,evaluation is made from the following aspects： mean,entropy,cross entropy,distortion of the correlation coefficients and the feasibility of the method is verified in this paper.