2009-056 – The CMU Multi-PIE Face Database

A close relationship exists between the advancement of face recognition algorithms and the availability of face databases varying factors that affect facial appearance in a controlled manner.The PIE database, collected at Carnegie Mellon University in 2000, has been very influential in advancing research in face recognition across pose and illumination. Despite its success the PIE database has several shortcomings: a limited number of subjects, a single recording session and only few expressions captured. To address these issues researchers at Carnegie Mellon University collected the Multi-PIE database. It contains 337 subjects, captured under 15 view points and 19 illumination conditions in four recording sessions for a total of more than 750,000 images. Labels are provided for a total of 6152 images of Multi-PIE. Illumination images were captured by recording 20 images within 0.7 seconds with flashes firing one by one. Due to the minimal subject movement between the different flash images the labels could be copied across all illumination conditions. These are AAM-style labels with between 39 and 68 feature points depending on the pose which were determined manually. Here is an illustration for the points labeled for the frontal view. CMU CTTEC CTTEC-Flintbox@andrew.cmu.edu 412-268-7393