From few to many: generative models for recognition under variable pose and illumination

By Athinodoros S. Georghiades, Peter N. Belhumeur, David J. Kriegman. Proceedings Fourth IEEE International Conference on Automatic Face and Gesture Recognition (Cat. No. PR00580), 2000

307 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/3a2faa145c5fe63ab906568a29fa4100220e03d9

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Appearance-based face recognition and light-fields (2004, background): “Algorithms have been proposedwhich dogeneralize acrosspose, for example [12], but this algorithm computes 3D head models using a gallery containing a large number of images per subject captured with controlled illumination variation.”
Face Recognition Under Varying Illumination Using Gradientfaces (2009, background): “In [10] and [11], the authors showed that face images with ﬁxed pose under varying illumination form a convex cone in the space of images, called illumination cone.”
Face recognition across pose: A review (2009, methodology): “All rights reserved. doi:10.1016/j.patcog.2009.04.017 a few promising methods have been proposed in tackling the problem of recognising faces in arbitrary poses, such as tied factor analysis (TFA) [63], 3D morphable model (3DMM) [14], eigen light-field (ELF) [33], illumination cone model (ICM) [30],”
Acquiring linear subspaces for face recognition under variable lighting (2005, background): “In the past few years, many appearance-based methods have been proposed to handle this problem, and new theoretical insights, as well as good recognition results, have been reported [1], [2], [3], [5], [7], [9].”
Deriving intrinsic images from image sequences (2001, background): “Figure 7: Results on one face from the Yale Face Database B [5].”
An Image Preprocessing Algorithm for Illumination Invariant Face Recognition (2003, methodology): “The Yale Face Database B [6] contains 5760 single light source images of 10 subjects each seen under 576 viewing conditions: 9 di(cid:11)erent poses and 64 illumination conditions.”
Joint Haar-like features for face detection (2005, methodology): “The test data, that is the original training data, contains difficult face examples with large change in illumination, such as images from Yale and XM2VTS databases.”
Clustering appearances of objects under varying illumination conditions (2003, background): “Each of these two subsets consists of 450 images with 45 images of each person acquired under varying light source directions ranging from frontal illumination to 70 degrees from frontal (See [11] for more details).”
Computing Gaussian Mixture Models with EM Using Equivalence Constraints (2003, methodology): “Comparative result s are given in Section 3, demonstrating the very significant advantage of our method over th two alternative constrained clustering algorithms using a number of data sets from the UC I repository and a large database of facial images [2].”
Lambertian reflectance and linear subspaces (2001, background): “In particular , we present an algorithm for recognition under varying pose and illumination that resembles [9], but works in a low-dimensional space that is derived analytically from a model.”