BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Using Deep Learning to Count Albatrosses from Space - Ellie Bowler
 \, University of East Anglia\, British Antarctic Survey
DTSTART:20200225T120000Z
DTEND:20200225T131500Z
UID:TALK137269@talks.cam.ac.uk
CONTACT:Jonathan Rosser
DESCRIPTION:Chair: Emily Shuckburgh\nAbstract: In this project we aim to a
 utomate the detection of Wandering Albatrosses in super high resolution sa
 tellite imagery (DigitalGlobe's WorldView-3)\, using state of the art deep
  learning approaches. We train a convolutional neural network to classify 
 and detect potential albatrosses\, achieving accuracy values of approximat
 ely 80% on the test set. By analysing the agreement between manually gener
 ated labels\, we show that these results are in fact in line with human pe
 rformance. We hope that the methods will streamline the analysis of WorldV
 iew-3 imagery\, allowing more frequent monitoring of a species which is of
  high conservation concern.
LOCATION:Bullard Lab\, Seminar Room
END:VEVENT
END:VCALENDAR