BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Deep Learning Techniques for Anomaly and Defect Detection Applicat
 ions
DTSTART:20190621T090000
DTEND:20190621T110000
DTSTAMP:20260406T220228Z
UID:4a594f411c266f3c210dcf7fb94f5a75ad05a84d15dcbc3a67956314
CATEGORIES:Conferences - Seminars
DESCRIPTION:David Honzatko\nEDIC candidacy exam\nExam president: Prof. Mar
 tin Jaggi\nThesis advisor: Prof. Pascal Fua\nCo-examiner: Dr. Martin Rajma
 n\n\nAbstract\nApplication of deep learning based computer vision algorith
 ms in industry is often problematic due to their need for a large annotate
 d datasets.\nParticularly demanding task is detection of defects and anoma
 lies in production.\nNot only is the annotation process expensive\, but al
 so it might be infeasible to obtain representative data for the annotation
 .\nIn a typical scenario\, there are hundreds of genuine samples and tens 
 of defective ones. For each of these there are multiple observations.\nIn 
 this work\, we investigate defect and anomaly detection architectures that
  could benefit from multiple observations\, capture the defining features 
 of the genuine samples\, and make use of the limited amount of annotations
  available.\n\nBackground papers\nAnomaly Detection: A Survey\, by Chandol
 a\, Varun\, Arindam Banerjee\, and Vipin Kumar. ACM Computing Surveys (CSU
 R) 41.3 (2009): 1-58.\nPixel recurrent neural networks\, by Van Den Oord
 \, Aäron\, Nal Kalchbrenner\, and Koray Kavukcuoglu. Proceedings of the 3
 3rd International Conference on  on Machine Learning-Volume 48. JMLR. or
 g\, 2016.\nDeformable Convolutional Networks\, by Dai\, Jifeng\, et al. 20
 17 IEEE International Conference on Computer Vision (ICCV). IEEE\, 2017.
 \n\n\n\n 
LOCATION:BC 329 https://plan.epfl.ch/?room==BC%20329
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR