BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Colloquium: Sampling and Reconstruction of High-Dimensional Vis
 ual Appearance
DTSTART:20190930T101500
DTEND:20190930T111500
DTSTAMP:20260527T163421Z
UID:ad1847dff3eb46418f724e03d9fa251ae7b4ea0cd0268f20a486853c
CATEGORIES:Conferences - Seminars
DESCRIPTION:By: Ravi Ramamoorthi - University of California San Diego\nVid
 eo of his talk\n\nAbstract:\nMany problems in computer graphics and comput
 er vision involve high-dimensional 3D-8D visual datasets. Real-time image 
 synthesis with changing lighting and view is often accomplished by pre-com
 puting the 6D light transport function (2 dimensions each for spatial posi
 tion\, incident lighting and viewing direction). A similar challenge is fo
 und in Monte Carlo rendering by sampling the paths of light transport\, re
 quiring integration over a 3D-8D space involving time\, lens effects for d
 epth of field\, pixel area\, soft shadows and global illumination. Realist
 ic image synthesis also often involves acquisition of appearance data or l
 ight transport from real-world objects\, and view synthesis from acquired 
 images\, involving 4D-6D functions.  In computer vision\, problems like l
 ighting insensitive facial recognition similarly involve understanding the
  space of appearance variation across lighting and view. Since hundreds of
  samples may be required in each dimension\, and the total size is exponen
 tial in the dimensionality\, brute force acquisition or precomputation is 
 often not even feasible.\n \nIn this talk\, we describe a signal-processi
 ng approach that exploits the coherence\, sparsity and inherent low-dimens
 ionality of the visual data\, to derive novel efficient sampling and recon
 struction algorithms. We describe a variety of new computational methods a
 nd applications\, from affine wavelet transforms for real-time rendering w
 ith area lights\, to space-time and space-angle frequency analysis for mot
 ion blur and global illumination\, to compressive sensing and deep learnin
 g for light transport acquisition and view synthesis.  The results have h
 ad substantial impact\, with sampling and denoising now adopted in all pro
 duction rendering\, and we will show recent results for relighting or chan
 ging view from only five or six images.  The work also points toward a un
 ified sampling theory applicable to many areas of signal processing\, comp
 uter graphics and computer vision.\n\nBio:\nRavi Ramamoorthi is the Ronald
  L. Graham professor of Computer Science at the University of California\,
  San Diego\, and founding Director of the UC San Diego Center for Visual C
 omputing. Prof. Ramamoorthi is an author of more than 150 refereed publica
 tions in computer graphics and computer vision\, including 75 at ACM SIGGR
 APH/TOG\, and has played a key role in building multi-faculty research gro
 ups that have been recognized as leaders in computer graphics and computer
  vision at Columbia\, Berkeley and UCSD. His research has been recognized 
 with a half-dozen early career awards\, including the ACM SIGGRAPH Signifi
 cant New Researcher Award in computer graphics in 2007\, and the President
 ial Early Career Award for Scientists and Engineers (PECASE) for his work 
 in physics-based computer vision in 2008.   He was elevated to IEEE and 
 ACM Fellow in 2017\, and inducted into the SIGGRAPH Academy in 2019.\n\nMo
 re information
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR