BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:From pixels to decision: image analysis at our service
DTSTART:20111128T171500
DTSTAMP:20260407T193319Z
UID:9a2b86e18f28fd1e70fd854b0a33831d8a13470eb1db41128ab36f22
CATEGORIES:Inaugural lectures - Honorary Lecture
DESCRIPTION:Prof. Jean-Philippe Thiran\nResearches in image analysis aim a
 t developing robust and efficient algorithms to automatically extract info
 rmation from digital images. Applications can be found in a large variety 
 of domains\, ranging from earth observation by satellite imaging or from v
 ideo surveillance to behavior analysis or medical imaging. In all these do
 mains\, the goal of image analysis is to develop methods able to convert r
 aw data\, i.e. the pixels\, into information. Making this conversion robus
 t and efficient is a challenge. To address it\, pixels are not enough. It 
 is necessary to include as much additional information as possible that is
  available on the problem at hand\, in the image analysis algorithms. This
  is the central topic of our research at the Signal Processing Lab 5 (LTS5
 ). During this lecture I will show how different types of prior informatio
 n\, such as object shape or behavior\, can be included in image analysis a
 lgorithms. I will show examples in diverse fields including facial express
 ion recognition\, brain imaging or pedestrian tracking\, and highlight our
  future research directions in this field.
LOCATION:polydome
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR