Much is known about how we recognize and identify human faces from static images. In the real world, however, we see people in motion. Our goal in this project, is to begin to study how we recognize people from moving images, under circumstances that are closer to those we encounter in the real world. There are two main components to this work: a.) the database project; and b.) the dynamic face recognition experiments.
The database project. To study human recognition under realistic circumstances, we first need a data base of pictures and movies of the same individuals that can be used in memory experiments. At present, we are gathering such images and movies. These include a set of nine static pictures taken from different viewpoints, a video of each person looking around a room, a video of the person speaking, and one or more videos of the person showing facial expressions (e.g., laughter, surprise, etc.). There are also videos of each person taken at a distance, which include natural gait information and gesturing. Full protocols and method for the collection of these images and videos are now online.
The experiments. We are beginning a series of experiments that are aimed at determining how movement in faces can contribute the quality or accuracy of human memory.
Joshua Harms, Sarah Snow, Anneliese West, Dana Roark, Charlie Lannen, Matt Pappas (not shown).