BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Image Classification Using a Background Prior - Daniel Keren\, Dep
 artment of Computer Science University of Haifa
DTSTART:20131202T100000Z
DTEND:20131202T110000Z
UID:TALK49133@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:A canonical problem in computer vision is category classificat
 ion (e.g. find all instances of human faces\, cars etc.\, in an image). Ty
 pically\, the input for training a classifier is a relatively small sample
  of positive examples\, and a much larger sample of negative examples\, wh
 ich in current applications can consist of images from thousands of catego
 ries.\nThe difficulty of the problem sharply increases with the dimension 
 and size of the negative example set. In this talk I will describe an effi
 cient and easy to apply classification algorithm\, which replaces the nega
 tive samples by a prior and then constructs a "hybrid" classifier which se
 parates the positive samples from this prior. The resulting classifier ach
 ieves an identical or better classification rate than SVM\, while requirin
 g far smaller memory and lower computational complexity to train and apply
 .\nWhile here it is applied to image classes\, the idea is general and can
  be applied to other domains.\nJoint work with Margarita Osadchy and Bella
  Fadida-Specktor.\nAn early version of this work was presented in ECCV 201
 2.\n
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station Road\, Cambridge
 \, CB1 2FB
END:VEVENT
END:VCALENDAR