Deep fisher faces
Harald Hanselmann, Shen Yan and Hermann Ney
Abstract
Most current state-of-the-art methods for unconstrained face recognition use deep
convolutional neural networks. Recently, it has been proposed to augment the typically
used softmax cross-entropy loss by adding a center loss trying to minimize the distance
between the face images and their class centers. In this work we further extend the center
(intra-class) loss with an inter-class loss reminiscent of the popular early face recognition
approach Fisherfaces. To this end we add a term that directly optimizes the distances of
the class centers appearing in a batch in dependence of the input images. We evaluate
the new loss on two popular databases for unconstrained face recognition, the Labeled
Faces in the Wild and the Youtube Faces database.
Session
Posters
Files
Paper (PDF)
DOI
10.5244/C.31.165
https://dx.doi.org/10.5244/C.31.165
Citation
Harald Hanselmann, Shen Yan and Hermann Ney. Deep fisher faces. In T.K. Kim, S. Zafeiriou, G. Brostow and K. Mikolajczyk, editors, Proceedings of the British Machine Vision Conference (BMVC), pages 165.1-165.11. BMVA Press, September 2017.
Bibtex
@inproceedings{BMVC2017_165,
title={Deep fisher faces},
author={Harald Hanselmann, Shen Yan and Hermann Ney},
year={2017},
month={September},
pages={165.1-165.11},
articleno={165},
numpages={11},
booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
publisher={BMVA Press},
editor={Tae-Kyun Kim, Stefanos Zafeiriou, Gabriel Brostow and Krystian Mikolajczyk},
doi={10.5244/C.31.165},
isbn={1-901725-60-X},
url={https://dx.doi.org/10.5244/C.31.165}
}