BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Delineating Dislocations
DTSTART:20180522T093000
DTEND:20180522T113000
DTSTAMP:20260407T152529Z
UID:2a5dc9a170a0a5b54bc9b96515496e2a1be5fb01811fe2eb380f2fbf
CATEGORIES:Conferences - Seminars
DESCRIPTION:Okan Altingövde\nEDIC candidacy exam\nExam president: Prof. C
 écile Hebert\nThesis advisor: Prof. Pacal Fua\nCo-examiner: Dr. Graham Kn
 ott\n\nAbstract\nAbstract—Dislocations in materials such as crystals car
 ry\nvaluable information which may be incorporated to characterize\ntheir 
 topology and interaction mechanisms. Therefore\, obtaining\naccurate recon
 struction of the 3D geometry of dislocations is\nthe key step for successf
 ully analyzing material characteristics.\nDelineation of dislocations is a
  challenging task due to ambiguities\nin corresponding structures between 
 images\, variations in data\ncaused by acquisition processes and limited r
 eal-world data. In\nthis proposal\, first we address three aspects\, which
  are detection\,\n3D reconstruction and domain adaptation\, of dislocation
 \ndelineation by discussing on related existing works: a deep\nlearning me
 thod for high accuracy edge detection\, several deep\nlearning approaches 
 for domain adaptations and a method to\nreveal geometry between three imag
 es of same scene taken from\ndifferent views. Finally\, we propose a resea
 rch plan which stands\non previous works on aforementioned three core part
 s while\neffectively combining them to achieve accurate delineation.\n\nBa
 ckground papers\nRicher Convolutional Features for Edge Detection\, by Liu
  Y\, et al.\nDeep-Learning Systems for Domain Adaptation in Computer Visio
 n\, by Venkateswara H.\, et al.\nWhat Can Two Images Tell Us About a Third
  One?\, by Faugeras O.\, Robert\, L.\n 
LOCATION:BC 329 https://plan.epfl.ch/?room==BC%20329
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR