Is average face recognized as the average

Is average face recognized as the average
复制标题

DOI:
10.1109/isspa.2007.4555339
复制
发表时间:
2007-02
期刊:
2007 9th International Symposium on Signal Processing and Its Applications
影响因子:
--
通讯作者:
N. P. Chandrasiri;R. Suzuki;N. Watanabe;Hiroyuki Yoshida;H. Yamada;H. Harashima
N. P. Chandrasiri;R. Suzuki;N. Watanabe;Hiroyuki Yoshida;H. Yamada;H. Harashima
中科院分区:
其他
文献类型:
--
作者:
N. P. Chandrasiri;R. Suzuki;N. Watanabe;Hiroyuki Yoshida;H. Yamada;H. Harashima

文献摘要

相似文献

在本文中,我们解决了一个非常基本的问题,即两张人脸图像的平均脸是否被人类识别为平均脸。首先,计算我们的刺激样本中每对图像的平均面孔。然后,我们进行了心理学实验,将平均人脸与用于计算的两张人脸图像进行比较。在实验中,参与者被指示在整体相似性方面进行比较,并选择与平均值更相似的图像。结果表明,参与者的反应是不均匀分布在原始图像刺激,这表明有一个显着的偏见,对一个平均的脸。基于主成分分析(PCA)构造特征脸空间。较高的相关性被发现之间的人和PCA的计算机识别。
In this paper, we address a very basic question of whether the average face of two face images is recognized as the average by a human. First, average faces were calculated for each of the pairs of images in our stimuli samples. Then, we conducted psychological experiments to compare an average face with two of the face images that were utilized for its calculation. In the experiment, participants were instructed to compare in terms of similarity as a whole and, select the image which is more similar to the average. Results showed that responses of the participants were nonuniformly-distributed across original image stimuli, suggesting that there was a significant bias on recognition of an average face. Eigenface space was constructed based on principal component analysis (PCA). Higher correlation was found between human and PCA based computer recognitions.