EIGENFEATURES AS INTERMEDIATE-LEVEL REPRESENTATIONS - THE CASE FOR PCA MODELS

Citation
H. Abdi et al., EIGENFEATURES AS INTERMEDIATE-LEVEL REPRESENTATIONS - THE CASE FOR PCA MODELS, Behavioral and brain sciences, 21(1), 1998, pp. 17
Citations number
15
Categorie Soggetti
Psychology, Biological",Neurosciences,"Behavioral Sciences
ISSN journal
0140525X
Volume
21
Issue
1
Year of publication
1998
Database
ISI
SICI code
0140-525X(1998)21:1<17:EAIR-T>2.0.ZU;2-B
Abstract
Eigenfeatures are created by the principal component approach (PCA) us ed on objects described by a low-level code (i.e., pixels, Gabor jets) . We suggest that eigenfeatures act like the flexible features describ ed by Schyns et al. They are particularly suited for face processing a nd give rise to class-specific effects such as the other-race effect. The PCA approach can be modified to accommodate top-down constraints.