BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Monocular 3D Shape Reconstruction Using Deep Neural Networks
DTSTART:20170627T090000
DTEND:20170627T110000
DTSTAMP:20260407T020947Z
UID:188fb0003f7c7f750fd137f39bf0f779282e71ce9b95c33ea78fb199
CATEGORIES:Conferences - Seminars
DESCRIPTION:Jan Bednarik\nEDIC candidacy exam\nExam president: Prof. Sabin
 e Süsstrunk\nThesis advisor: Prof. Pascal Fua\nThesis co-advisor: Dr. Mat
 hieu Salzmann\nCo-examiner: Prof. Mark Pauly\n\nAbstract\nMonocular recons
 truction of non-rigid surfaces is a long-standing computer vision problem 
 which has so far been addressed by carefully selecting image features and 
 explicitly imposing physical constraints. We attempt to tackle the problem
  by learning to predict the shape directly from data. To improve on the re
 sults obtained from straight-forward regressor we explore the domain of re
 presentation learning with focus on adversarial setting which might help p
 roduce more realistic shape predictions. Then\, we discuss the approaches 
 to estimate the scene lighting which would further help to reason about th
 e shape using shading clues. Finally\, we summarize preliminary results on
  both the synthetic and real dataset and we discuss future directions.\n\n
 Background papers\nDensity estimation using real NVP\, by L. Dinh et al I
 CLR\, 2016.\nOn the relationship between radiance and irradiance: determin
 ing the illumination from images of a convex lambertian object\, by R. Ram
 amoorthi and P. Hanrahan. JOSA A\, 2001.\nInfoGAN: Interpretable represent
 ation learning by information maximizing generative adversarial nets\, by
  X. Chen et al. NIPS\, 2016.
LOCATION:BC 329 https://plan.epfl.ch/?room==BC%20329
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR