基于奇异值的具有年龄变化的人脸识别

Face recognition with ageing based on SVD

为了增强现有人脸识别算法对年龄变化的鲁棒性,提出了一种新的基于奇异值分解(SVD)和嵌入式隐马尔可夫模型(EHMM)的人脸识别方法。先选取整幅人脸图像的奇异值作为基本特征向量,然后建立年龄函数,对奇异值特征进行修正,再根据得到的年龄函数,对人脸图像进行重建,提取改进后的奇异值特征作为观察序列,送入EHMM中进行分类识别。实验结果表明这种方法能够提高具有年龄变化的人脸识别效率。

A new method of face recognition based on singular value decomposition （SVD） and embedded hidden markov model （EHMM） is presented to enhance the ability of the current face recognition algorithms robust to ageing. First the SVD vectors of the total image as basic feature vectors are selected, and second the age function and adjusts the vectors accordingly are built. Then the image can be reshaped based on the function above, and the new adjusted feature vectors is extracted to be put into the EHMM classification as inputs for classifying and recognizing. The experimental results show that this method can improve the recognition rate of face images with aging.