UTCS Colloquia/AI - Derek Hoiem/University of Illinois at Urbana-Champaign, "Representing and Inferring the 3D Layout of Rooms", ACES 2.302

Contact Name: 
Jenna Whitney
Apr 6, 2012 11:00am - 12:00pm

There is a sign-up schedule for this event that can be found at


Type o

f Talk: UTCS Colloquia/AI

Speaker/Affiliation: Derek Hoiem/University

of Illinois at Urbana-Champaign

Talk Audience: UTCS Faculty, Graduate
Students, Undergraduate Students and Outside Interested Parties


/Time: Friday, April 6, 2012, 11:00 am

Location: ACES 2.302


st: Kristen Grauman

Talk Title: Representing and Inferring the 3D Layo

ut of Rooms

We humans spend much of our time working,

playing, and sleeping in rooms and can operate effectively in them. But co

mputers have difficulty interpreting rooms because many important surfaces

are hidden and nearby objects exhibit confusing perspective effects. For co

mputers to interpret, navigate, or interact in rooms, they need better r

epresentations of space. I will describe our efforts to represent and infer
3D layout of rooms. I will show how we can use structural priors to detect
hidden boundaries and to create perspective-robust models of object appear

ance. I''ll present results for inferring the 3D layout of walls and furnit

ure from one image. I''ll also present an application to inserting 3D objec

ts into an image and relighting them (including cool videos). Finally, I''

ll touch on our most recent work to organize an RGB-depth image into surfac

es and objects with physical support relations and will discuss the most pr

essing future directions.

Derek Hoiem is an assistant professor
in Computer Science at the University of Illinois at Urbana-Champaign (UIU

C). Before joining the UIUC faculty in 2009, Derek completed his Ph.D. in

Robotics at Carnegie Mellon University in 2007 and was a postdoctoral fello

w at the Beckman Institute from 2007-2008. Derekā€šs research interests foc

us on 3D scene interpretation and object recognition. His work has been rec

ognized with a CVPR 2006 Best Paper award, a 2008 ACM Doctoral Dissertatio

n Award honorable mention, and a 2011 NSF CAREER award.