Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection
Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection
复制标题
DOI：
10.1007/bfb0015522
复制
发表时间：
1996-04
期刊：
影响因子：
--
通讯作者：
P. Belhumeur;J. Hespanha;D. Kriegman
中科院分区：
文献类型：
--
作者：
P. Belhumeur;J. Hespanha;D. Kriegman
我们开发了一种人脸识别算法，它是不敏感的照明方向和面部表情的变化。以模式分类的方法，我们认为每个像素在图像中的一个坐标在一个高维空间。我们利用的观察，一个特定的脸在不同的照明方向下的图像位于一个3-D线性子空间的高维特征空间-如果面对的是一个朗伯表面没有自阴影。然而，由于人脸不是真正的朗伯曲面，并且确实会产生自阴影，因此图像将偏离此线性子空间。而不是明确地建模这种偏差，我们将图像投影到一个子空间的方式，折扣的脸与大偏差的区域。我们的投影方法是基于费舍尔的线性判别，并产生良好的分离类在低维子空间，即使在光照和面部表情的严重变化。特征脸技术是另一种基于将图像空间线性投影到低维子空间的方法，具有类似的计算要求。然而，大量的实验结果表明，所提出的“Fisherface”方法的错误率显着低于那些在同一数据库上进行测试时的特征脸技术。
We develop a face recognition algorithm which is insensitive to gross variation in lighting direction and facial expression. Taking a pattern classification approach, we consider each pixel in an image as a coordinate in a high-dimensional space. We take advantage of the observation that the images of a particular face under varying illumination direction lie in a 3-D linear subspace of the high dimensional feature space — if the face is a Lambertian surface without self-shadowing. However, since faces are not truly Lambertian surfaces and do indeed produce self-shadowing, images will deviate from this linear subspace. Rather than explicitly modeling this deviation, we project the image into a subspace in a manner which discounts those regions of the face with large deviation. Our projection method is based on Fisher's Linear Discriminant and produces well separated classes in a low-dimensional subspace even under severe variation in lighting and facial expressions. The Eigenface technique, another method based on linearly projecting the image space to a low dimensional subspace, has similar computational requirements. Yet, extensive experimental results demonstrate that the proposed “Fisherface” method has error rates that are significantly lower than those of the Eigenface technique when tested on the same database.