Despite recent major advances, face detection is still a challenge in difficult conditions, when the face is not frontal and/or the illumination conditions are adverse.
Recently, a new database, called MOBIO, has been publicly released. The MOBIO database is a large and challenging biometric database. It contains about 28'800 videos of 150 participants captured in challenging real-world conditions on a mobile phone camera over a one and a half year period.
For this challenge, we propose a face detection task and optionally an eye localization task using the MOBIO database. Instead of processing the videos, one image was extracted from each of them and was manually annotated with eye locations. These manual locations will be used to assess the performance of both face detection and eye localization using a standard metric such as the Jesorsky measure.
- Dr. Sebastien Marcel, Idiap Research Institute, Switzerland