In this page, we provide a list of Reproducible Research publications and other examples that build on top of Bob.
If you want your publication to be listed here, please contact us.
These publications use the new Bob framework and should work out-of-the-box.
Please, let us know in case you have trouble installing or executing any of the papers below.
We compare a list of face recognition algorithms' performances on different aspects of face recognition, such as facial expression, occlusion, pose and the availability of video data.
We perform gender classification using a boosted strong classifier build out of weak look-up-table based classifiers, which are build from overlapping multi-block LBP codes.
We test, how several face recognition algorithms perform, when the images are mis-aligned due to an imprecise eye localization.
These publications rely on an older version of Bob, which need Bob v1 to be installed.
We perform score calibration, a technique adopted from the speaker recognition domain, to perform score calibration using face recognition scores.
We use score calibration techniques to fuse the recognition scores from several speaker and face recognition algorithms and show that the fusion outperforms each single algorithm by far.
We introduce the FaceRecLib (which in Bob v2 is split into the bob.bio packages), a tool to provide a fair comparison of face recognition algorithms.