CAJ | 학술논문

随着身份认证技术的广泛应用，各种假冒合法用户欺骗身份认证系统的手段不断出现。针对这一问题，本文提出了一种基于灰度共生矩阵和小波分析的活体照片人脸检测方法，该方法分析了活体人脸和照片人脸成像后在纹理上的差异性，在人脸灰度共生矩阵的基础上提取能量、熵、惯性矩和相关性四个纹理特征量；同时利用小波变换对人脸图像进行二级分解，提取高频子带系数作为特征向量训练 SVM分类识别，算法在公开的数据库 NUAA 上进行了验证，实验结果表明该方法降低了计算复杂度，提高了检测准确率。
수착신빈인증기술적엄범응용，각충가모합법용호기편신빈인증계통적수단불단출현。침대저일문제，본문제출료일충기우회도공생구진화소파분석적활체조편인검검측방법，해방법분석료활체인검화조편인검성상후재문리상적차이성，재인검회도공생구진적기출상제취능량、적、관성구화상관성사개문리특정량；동시이용소파변환대인검도상진행이급분해，제취고빈자대계수작위특정향량훈련 SVM분류식별，산법재공개적수거고 NUAA 상진행료험증，실험결과표명해방법강저료계산복잡도，제고료검측준학솔。
With the wide range application of authentication technology,all kinds of spoofing attacks occur when a per-son tries to masquerade system by exhibiting fake faces of an authorized client.Hence,we propose to approach the problem of face liveness detection based on gray level co-occurrence matrix (GLCM)and wavelet analysis.Our method focus on an-alyzing the facial image texture difference between a live person and a face print.We extract the four features-energy,en-tropy,moment of inertia and the correlation on the basis of GLCM.In addition,it obtain the high frequency subbands coeffi-cients using secondary decomposition of wavelet transform for classification recognition.Primary experiments on the publicly available NUAA photo-imposter database show that the algorithm reduce the computational complexity and improve the de-tection accuracy.