摘要: 在进行人脸识别时,光照、表情、角度等因素的影响会大幅增加数据计算的时空复杂度。为此,提出一种新的图像外观统计模型,在动态形状模型中引入灰度共生矩阵(GLCM),通过计算图像形状对齐情况下的GLCM,建立半动态外观模型。基于ORL人脸数据库的实验结果表明,该模型相比动态外观模型,识别准确率更高,速度更快。
关键词:
半动态外观模型,
动态形状模型,
灰度共生矩阵,
动态外观模型,
人脸识别
Abstract: The influence of the illumination, expression and the angle and so on may enhance the time complexity and space complexity of the data computation greatly during the face recognition, so this paper proposes a new statistical model for image appearance. Active Shape Model (ASM) is introduced in Grey Level Co-occurrence Matrix(GLCM). By calculating the GLCM under the shape of the image alignment, Semi-active Appearance Model(SAAM) is established. Experiments on the standard ORL face database show that compared with ASM, the model gains higher recognition rate and speed.
Key words:
Semi-active Appearance Model(SAAM),
Active Shape Model(ASM),
Grey Level Co-occurrence Matrix(GLCM),
Active Appearance Model(AAM),
face recognition
中图分类号:
杨占栋, 解梅. 基于半动态外观模型的人脸识别[J]. 计算机工程, 2011, 37(24): 150-151.
YANG Tie-Dong, JIE Mei. Face Recognition Based on Semi-active Appearance Model[J]. Computer Engineering, 2011, 37(24): 150-151.