Available Datasets

English Images

This dataset contains all the image data from the Controlled, Degraded and Adverse recording sessions for the English portion of the database, i.e. a total of 6,240 images. The dataset allows you to perform experiments according to the full BANCA protocol. It also includes an additional 300 images taken from the world model set.

English Audio
The audio extracted from the sequences.
English Sequences
The associated audio-video sequences of the BANCA English Images. There are 1308 sequences in total