facereclib 1.1.1

Databases

To achieve this goal, interfaces to many publicly available facial image databases are contained, and default evaluation protocols are defined, e.g.:

• Face Recognition Grand Challenge version 2 [http://www.nist.gov/itl/iad/ig/frgc.cfm]

• The Good, The Bad and the Ugly [http://www.nist.gov/itl/iad/ig/focs.cfm]

• Labeled Faces in the Wild [http://vis-www.cs.umass.edu/lfw]

• Multi-PIE [http://www.multipie.org]

• SCface [http://www.scface.org]

• MOBIO [http://www.idiap.ch/dataset/mobio]

• BANCA [http://www.ee.surrey.ac.uk/CVSSP/banca]

• CAS-PEAL [http://www.jdl.ac.cn/peal/index.html]

• AR face database [http://www2.ece.ohio-state.edu/~aleix/ARdatabase.html]

• XM2VTS [http://www.ee.surrey.ac.uk/CVSSP/xm2vtsdb]

• The AT&T database of faces (formerly known as ORL) [http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html]

Algorithms

Together with that, a broad variety of traditional and state-of-the-art face recognition algorithms such as:

• Eigenfaces [TP91]

• Linear Discriminant Analysis [ZKC+98]

• Probabilistic Linear Discriminant Analysis [ESM+13]

• Local Gabor Binary Pattern Histogram Sequences [ZSG+05]

• Graph Matching [GHW12]

• Gaussian Mixture Modeling [MM09]

• Inter-Session Variability Modeling [WMM+11]

• Bayesian Intrapersonal/Extrapersonal Classifier [MWP98]

is provided. Furthermore, tools to evaluate the results can easily be used to create scientific plots, and interfaces to run experiments using parallel processes or an SGE grid are provided.

Extensions

On top of these already pre-coded algorithms, the FaceRecLib provides an easy Python interface for implementing new image preprocessors, feature types, face recognition algorithms or database interfaces, which directly integrate into the face recognition experiment. Hence, after a short period of coding, researchers can compare their new invention directly with already existing algorithms in a fair manner.

As a small example, we provide wrapper classes for the CSU face recognition resources [http://www.cs.colostate.edu/facerec] in the xfacereclib.extension.CSU package:

• Local Region PCA [PBD+11]

• LDA-IR (a.k.a. CohortLDA) [LBP+12]

To see how easy it is to use these tools to generate a publication, please have a look at the source code xfacereclib.paper.BeFIT2012, which was finally published in [M. Günther, R. Wallace and S. Marcel. “An Open Source Framework for Standardized Comparisons of Face Recognition Algorithms”. Computer Vision - ECCV 2012. Workshops and Demonstrations, LNCS, 7585, 547-556, 2012.].

Finally, all parts of the FaceRecLib are well documented and thoroughly tested to assure usability, stability and comparability.

References

Bob

You will need a copy of Bob in version 1.2.0 to run the algorithms. Please download Bob from its webpage.

After downloading, you should go to the console and write:

$ python bootstrap.py
$ bin/buildout
$ bin/sphinx-build docs sphinx

This will download all required packages and install them locally. If you don’t want all the database packages to be downloaded, please remove the xbob.db.[database] lines from the eggs section of the file buildout.cfg in the main directory before calling the three commands above.

Now, you can open the documentation by typing:

$ firefox sphinx/index.html

and read further instructions on how to use this library.

The CSU Face Recognition Resources

Two open source algorithms are provided by the CSU Face Recognition Resources, namely the LRPCA and the LDA-IR (a.k.a. CohortLDA) algorithm. For these algorithms, optional wrapper classes are provided in the xfacereclib.extension.CSU satellite package. By default, this package is disabled. To enable them, please call:

$ bin/buildout -c buildout-with-csu.cfg

after downloading and patching the CSU resources, and updating the sources-dir in the buildout-with-csu.cfg file – as explained in xfacereclib.extension.CSU.