The BANCA database is a new large, realistic and challenging multi-modal database intended for training and testing multi-modal verification systems. The BANCA database was captured in four European languages in two modalities (face and voice). For recording, both high and low quality microphones and cameras were used. The subjects were recorded in three different scenarios, controlled, degraded and adverse over 12 different sessions spanning three months. In total 208 people were captured, half men and half women.
Associated with the database is the BANCA protocol. The protocol defines which sets of data to use for training, evaluation and testing. Performing experiments according to the protocol allows institutions to easily compare their results to others. Two face verification competitions on the images from the BANCA database and associated protocol are being held in the year 2004. The first is being held in conjunction with ICBA and the second in conjunction with ICPR 2004.
Through this web-site portions of the BANCA database are being made available to the research community. As more of the data becomes available it will be released here. Presently, the complete set of English images is available.
The BANCA database offers the research community the opportunity to test their multi-modal verification algorithms on a large, realistic and challenging database. It is hoped that this database and protocol will become a standard, like the XM2VTS database, which enables institutions to easily compare the performance of their own algorithms to others.
|The BANCA Database||The XM2VTS Database|