We describe a new face database for developing machine learning approaches to face perception. In particular we intend the database to be used for developing systems for facial identity and expression recognition. Face images were collected from a variety of sources and subsumed into a common set of files. A face detector was run on each image and then images were centered and cropped based on the output of the face detector. Many of the images came from third-party sources including datasets that were purchased, required consent, and/or were available freely online.