BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Scene Decomposition and Relighting from Image Collections in Neura
 l Rendering
DTSTART:20220823T140000
DTEND:20220823T160000
DTSTAMP:20260407T230640Z
UID:38cd4c15f5a4231eab9d9a0f9f6abb69b13d929b94f71d11d58ee6e3
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dongqing Wang\nEDIC candidacy exam\nExam president: Prof. Mark
  Pauly\nThesis advisor: Prof. Sabine Süsstrunk\nCo-examiner: Prof. Ronan 
 Boulic\n\nAbstract\nThe focus of our research is to generate controllable 
 photo-realistic images of real-world scenes from their existing sparse obs
 ervations\, i.e.\, the inverse rendering problem. The approaches we focus 
 on are those through neural rendering\, utilizing neural network to decomp
 ose the scene\, learn its physical properties and render with novel lighti
 ng condition. In this proposal\, we discuss three papers and how they rela
 te to our research topic. We first look at NeRF's simple framework represe
 nting 3D scenes as volumetric radiance field for view synthesis\; Then we 
 look at NeRD which modifies NeRF to allow scene decomposition for illumina
 tion\, geometry\, surface reflectance\, etc.\, for relighting\; we lastly 
 present PhySG using signed distance functions for scene geometry addressin
 g drawback of previous methods. Finally\, we discuss our proposed solution
  for the problem as well as possible future research focus.\n\nBackground 
 papers\n\n1. NeRF: Representing Scenes as Neural Radiance Fields for View
  Synthesis. \n\n2. NeRD: Neural Reflectance Decomposition from Image Co
 llections\n\n3. PhySG: Inverse Rendering with Spherical Gaussians for Phy
 sics-based Material Editing and Relighting.\n\n\n 
LOCATION:BC 329 https://plan.epfl.ch/?room==BC%20329
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR