UTCS Colloquium-Devi Parikh/Carnegie Mellon University: "The Role of Context in Image Understanding: When, For What, and How?" ACES 3.408, Tuesday, April 7, 2009 11:00 a.m.

Contact Name: 
Jenna Whitney
Date: 
Apr 7, 2009 11:00am - 12:00pm

There is a signup schedule for this event (UT EID required).
Type of T

alk: UTCS ColloquiumSpeaker/Affiliation: Devi Parikh/Carnegie

Mellon UniversityDate/Time: Tuesday, April 7, 2009 11:00 a.m.

Location: ACES 3.408Host: Kristen GraumanTalk Title:
"The Role of Context in Image Understanding: When, For What, and

How?"Talk Abstract:A key problem in computer vision

is image understanding, which we define as the task of recognizing every o

bject in the scene, and perhaps the scene category itself. Traditionally,
object recognition has been accomplished by considering only the informati

on within the object to be recognized. Incorporating contextual information

, i.e., information outside the boundaries of the object, for enhanced r

ecognition has received significant attention in recent works. In this talk

, we take a closer look at the role of context. Specifically, we ask thre

e questions. First: When is context really helpful? We show, through compu

ter vision experiments as well as human studies, that context provides imp

rovements in recognition performances only when the appearance information

is weak (such as in low resolution images or in the presence of occlusion).
Second: For what tasks can contextual information be leveraged? We show th

at apart from high-level tasks of object recognition and detection, contex

tual information can be effectively leveraged for low level tasks as well,
such as identifying salient or representative patches in an image. Lastly

, How can context be learnt? Or alternatively, how much contextual informa

tion can be extracted in an unsupervised manner? We propose a unified hiera

rchical representation for contextual interactions or spatial patterns amon

g visual entities at all levels, from low-level features to parts of objec

ts, objects, groups of objects and ultimately the entire scene. We presen

t results of our approach on a variety of datasets such as object categorie

s, street scenes and natural scene images.