BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Crowd Density Estimation With Convolutional Neural Networks
DTSTART:20180611T090000
DTEND:20180611T110000
DTSTAMP:20260501T134030Z
UID:68f858acf28e28125fea7656f049600097030f567b0f1848624b908b
CATEGORIES:Conferences - Seminars
DESCRIPTION:Weizhe Liu\nEDIC candidacy exam\nExam president: Prof. Bertran
 d Merminod\nThesis advisor: Prof. Pascal Fua\nCo-examiner: Dr. Ronan Bouli
 c\n\nAbstract\nAbstract—Crowd analysis has attracted significant attenti
 on\nin past years due to its increasing application in video surveillance\
 ,\ntraffic control and emergency management. Like any\nother problems in c
 omputer vision\, crowd analysis comes with\nmany challenges such as occlus
 ions\, high clutter\, non-uniform\ndistribution of people and time-consumi
 ng annotating dataset.\nScale and perspective make the problem extremely d
 ifficult to\nsolve. Our research goal is to address these challenges espec
 ially\nperspective distortion.\nIn this proposal\, we discuss three existi
 ng work and how they\nrelate to our research. We first explain how researc
 hers deal\nwith this problem without using convolutional neural networks.\
 nThen we introduce a survey of recent CNN-based methods\nin addressing abo
 ve mentioned problems. Finally we discuss\nan example of enforcing tempora
 l consistency into video-based\ncrowd counting and compare these methods w
 ith our own ideas.\n\nBackground papers\nSegmentation and tracking of mult
 iple humans in crowded environments\, by Zhao\, Tao\, Ram Nevatia\, and Bo
  Wu  IEEE transactions on pattern analysis and machine intelligence 30.7
  (2008): 1198-1211.\nA survey of recent advances in cnn-based single image
  crowd counting and density estimation\, by Sindagi\, Vishwanath A.\, and 
 Vishal M. Patel. Pattern Recognition Letters (2017).\nSpatiotemporal mode
 ling for crowd counting in videos by  Xiong\, Feng\, Xingjian Shi\, and D
 it-Yan Yeung. 2017 IEEE International Conference on Computer Vision (ICCV)
 . IEEE\, 2017.
LOCATION:BC 229 https://plan.epfl.ch/?room==BC%20229
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR