tailieunhanh - The CMU Pose, Illumination, and Expression (PIE) Database

In this section, we study operations that, in one way or another, process their data sequentially. Examples include searching on one or more unindexed attributes, creation of bitmaps based on selection conditions, and scalar aggrega- tion. Grouped aggregation can be achieved by a combina- tion of sorting followed by one scalar aggregation per group. Scan-like operations are also used as components of more complex operations that we will discuss in Section 4 and Section 5. In practice, we need to consider inputs that may be un- aligned, or where the number of elements may not be a multiple of S. Special pieces of code can be used for the rst and last. | Appeared in the 2002 International Conference on Automatic Face and Gesture Recognition The CMU Pose Illumination and Expression PIE Database Terence Sim Simon Baker and Maan Bsat The Robotics Institute Carnegie Mellon University 5000 Forbes Avenue Pittsburgh PA 15213 Abstract Between October 2000 and December 2000 we collected a database of over 40 000 facial images of 68 people. Using the CMU 3D Room we imaged each person across 13 different poses under 43 different illumination conditions and with 4 different expressions. We call this database the CMU Pose Illumination and Expression PIE database. In this paper we describe the imaging hardware the collection procedure the organization of the database several potential uses of the database and how to obtain the database. 1 Introduction People look very different depending on a number of factors. Perhaps the three most significant factors are 1 the pose . the angle at which you look at them 2 the illumination conditions at the time and 3 their facial expression . whether or not they are smiling etc. Although several other face databases exist with a large number of subjects Philips et al. 1997 and with significant pose and illuminationvariation Georghiades et al. 2000 we felt that there was still a need for a database consisting of a fairly large number of subjects each imaged a large number of times from several different poses under significant illumination variation and with a variety of facial expressions. Between October 2000 and December 2000 we collected such a database consisting of over 40 000 images of 68 subjects. The total size of the database is about 40GB. We call this database the CMU Pose Illumination and Expression PIE database. To obtain a wide variation across pose we used 13 cameras in the CMU 3D Room Kanade et al. 1998 . To obtain significant illumination variation we augmented the 3D Room with a flash system similar to the one constructed by Athinodoros Georghiades Peter Bel-humeur and