MultiPie Dataset is a multi-view face image dataset, which is mainly used for identity verification. It is an alternative to the PIE database. It contains 337 subjects captured under 15 viewpoints, covering a total of more than 750,000 images. The dataset was released by Carnegie Mellon University in 2009.

The PIE database was collected by Carnegie Mellon University in 2000 and has been influential in advancing research in pose and illumination face recognition. Despite its success, the PIE database has the following shortcomings: a limited number of subjects, a single recording session, and few captured expressions.