BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:identity variables for face recognition: from distance based metho
 ds to probabilistic inference - Simon Prince\, University College London
DTSTART:20070605T140000Z
DTEND:20070605T150000Z
UID:TALK7637@talks.cam.ac.uk
CONTACT:Oliver Williams
DESCRIPTION:Many face recognition algorithms use ``distance-based`` method
 s: feature vectors are extracted from each face and distances in feature s
 pace are compared to determine matches. In this paper we argue for a funda
 mentally different approach. We consider each image as having been generat
 ed from an underlying cause (a latent identity variable\, or LIV). In reco
 gnition we evaluate the probability that two faces have the same underlyin
 g cause. Since image generation is noisy\, we can never be exactly certain
  what this cause was\, so we integrate (marginalize) over all possible cau
 ses. We present examples of identification and verification and show that 
 the LIV approach outperforms equivalent distance-based algorithms. Moreove
 r\, other advantages include: (i) a natural approach to changes in pose an
 d lighting (ii) the ability to implement novel algorithms that have no dis
 tance-based equivalent (iii) a principled way to combine multiple observat
 ions and prior information. Finally we demonstrate how this framework can 
 be applied to the more general problem of clustering face images. 
LOCATION:Small public lecture room\, Microsoft Research Ltd\, 7 J J Thomso
 n Avenue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR