Eigenface算法与EBGM算法的适应性比较

Eigenface算法和EBGM算法是人脸识别的两种重要算法。前者基于图像的整体特征,后者通过Gabor变换提取图像的局部特征。在实际应用中,光照的变化、人物表情的变化和物体对人脸的遮盖等因素造成了人脸识别的困难。文章对上述两种算法在这些变化因素下的识别性能进行了研究和比较。实验结果表明EBGM算法对环境变化具有更好的适应性,能够在小样本条件下获得良好的识别能力。而Eigenface算法对环境变化较为敏感,需要大量的训练样本来保证识别效果。

The Comparison of Adaptability for Eigenface and EBGM

Eigenface and EBGM are the main algorithms for face recognition.Eigenface is based on the features of entire image and EBGM uses Gabor transform to extract local features.In practice,changes of illumination,face expressions and facial details bring much trouble for face recognition.In this paper,we have compared the adaptability of Eigenface and EBGM under varying conditions.Experiments show that EBGM,which only needs few training samples, has better adaptability than Eigenface,which needs much more samples to ensure the discriminating power.