<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii">
<meta name=Generator content="Microsoft Word 12 (filtered medium)">
<style>
<!--
/* Font Definitions */
@font-face
{font-family:"Cambria Math";
panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
{font-family:Consolas;
panose-1:2 11 6 9 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
margin-bottom:.0001pt;
font-size:11.0pt;
font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
{mso-style-priority:99;
color:purple;
text-decoration:underline;}
p.MsoPlainText, li.MsoPlainText, div.MsoPlainText
{mso-style-priority:99;
mso-style-link:"Plain Text Char";
margin:0in;
margin-bottom:.0001pt;
font-size:10.5pt;
font-family:Consolas;}
span.EmailStyle17
{mso-style-type:personal-compose;
font-family:"Calibri","sans-serif";
color:windowtext;}
span.PlainTextChar
{mso-style-name:"Plain Text Char";
mso-style-priority:99;
mso-style-link:"Plain Text";
font-family:Consolas;}
.MsoChpDefault
{mso-style-type:export-only;}
@page WordSection1
{size:8.5in 11.0in;
margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
{page:WordSection1;}
-->
</style>
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=WordSection1>
<p class=MsoPlainText>Here’s a reminder of Sven Dickinson’s talk on
Wednesday....<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>UCSB COMPUTER SCIENCE DEPARTMENT PRESENTS:<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>Wednesday, October 6, 2010<o:p></o:p></p>
<p class=MsoPlainText>11:00 AM - 12:00 PM<o:p></o:p></p>
<p class=MsoPlainText>Computer Science Conference Room, Harold Frank Hall Rm.
1132<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>HOST: Matthew Turk<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>SPEAKER: Sven Dickinson<o:p></o:p></p>
<p class=MsoPlainText>Department of Computer Science<o:p></o:p></p>
<p class=MsoPlainText>University of Toronto<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>Title: The Role of Intermediate Shape Priors in
Perceptual Grouping and <o:p></o:p></p>
<p class=MsoPlainText>Image Abstraction<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>Abstract:<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>Perceptual grouping played a prominent role in support of
early object <o:p></o:p></p>
<p class=MsoPlainText>recognition systems, which typically took an input image
and a database <o:p></o:p></p>
<p class=MsoPlainText>of shape models and identified which of the models was
visible in the <o:p></o:p></p>
<p class=MsoPlainText>image. When the database was large, local features were
not <o:p></o:p></p>
<p class=MsoPlainText>sufficiently distinctive to prune down the space of models
to a <o:p></o:p></p>
<p class=MsoPlainText>manageable number that could be verified. However, when
causally <o:p></o:p></p>
<p class=MsoPlainText>related shape features were grouped, using
intermediate-level shape <o:p></o:p></p>
<p class=MsoPlainText>priors, e.g., cotermination, symmetry, and compactness,
they formed <o:p></o:p></p>
<p class=MsoPlainText>effective shape indices and allowed databases to grow in
size. In <o:p></o:p></p>
<p class=MsoPlainText>recent years, the recognition (categorization) community
has focused on <o:p></o:p></p>
<p class=MsoPlainText>the object detection problem, in which the input image is
searched for a <o:p></o:p></p>
<p class=MsoPlainText>specific target object. Since indexing is not required
to select the <o:p></o:p></p>
<p class=MsoPlainText>target model, perceptual grouping is not required to
construct a <o:p></o:p></p>
<p class=MsoPlainText>discriminative shape index; the existence of a much
stronger <o:p></o:p></p>
<p class=MsoPlainText>object-level shape prior precludes the need for a weaker <o:p></o:p></p>
<p class=MsoPlainText>intermediate-level shape prior. As a result, perceptual
grouping <o:p></o:p></p>
<p class=MsoPlainText>activity at our major conferences has diminished.
However, there are <o:p></o:p></p>
<p class=MsoPlainText>clear signs that the recognition community is moving from
appearance <o:p></o:p></p>
<p class=MsoPlainText>back to shape, and from detection back to unexpected
object <o:p></o:p></p>
<p class=MsoPlainText>recognition. Shape-based perceptual grouping will play a
critical role <o:p></o:p></p>
<p class=MsoPlainText>in facilitating this transition. But while causally
related features <o:p></o:p></p>
<p class=MsoPlainText>must be grouped, they also need to be abstracted before
they can be <o:p></o:p></p>
<p class=MsoPlainText>matched to categorical models. In this talk, I will
describe our <o:p></o:p></p>
<p class=MsoPlainText>recent progress on the use of intermediate shape priors
in segmenting, <o:p></o:p></p>
<p class=MsoPlainText>grouping, and abstracting shape features. Specifically,
I will describe <o:p></o:p></p>
<p class=MsoPlainText>the use of symmetry and non-accidental attachment to
detect and group <o:p></o:p></p>
<p class=MsoPlainText>symmetric parts, the use of closure to separate figure
from background, <o:p></o:p></p>
<p class=MsoPlainText>and the use of a vocabulary of simple shape models to
group and abstract <o:p></o:p></p>
<p class=MsoPlainText>image contours.<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>Bio:<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>Sven Dickinson received the B.A.Sc. degree in Systems
Design Engineering <o:p></o:p></p>
<p class=MsoPlainText>from the University of Waterloo, in 1983, and the M.S.
and Ph.D. degrees <o:p></o:p></p>
<p class=MsoPlainText>in Computer Science from the University of Maryland, in
1988 and 1991, <o:p></o:p></p>
<p class=MsoPlainText>respectively. He is currently Professor and Chair of the
Department of <o:p></o:p></p>
<p class=MsoPlainText>Computer Science at the University of Toronto, where he
has also served <o:p></o:p></p>
<p class=MsoPlainText>as Acting Chair (2008-2009), Vice Chair (2003-2006), and
Associate <o:p></o:p></p>
<p class=MsoPlainText>Professor (2000-2007). From 1995-2000, he was an
Assistant Professor of <o:p></o:p></p>
<p class=MsoPlainText>Computer Science at Rutgers University, where he also
held a joint <o:p></o:p></p>
<p class=MsoPlainText>appointment in the Rutgers Center for Cognitive Science
(RuCCS) and <o:p></o:p></p>
<p class=MsoPlainText>membership in the Center for Discrete Mathematics and
Theoretical <o:p></o:p></p>
<p class=MsoPlainText>Computer Science (DIMACS). From 1994-1995, he was a
Research Assistant <o:p></o:p></p>
<p class=MsoPlainText>Professor in the Rutgers Center for Cognitive Science,
and from <o:p></o:p></p>
<p class=MsoPlainText>1991-1994, a Research Associate at the Artificial
Intelligence <o:p></o:p></p>
<p class=MsoPlainText>Laboratory, University of Toronto. He has held
affiliations with the MIT <o:p></o:p></p>
<p class=MsoPlainText>Media Laboratory (Visiting Scientist, 1992-1994), the
University of <o:p></o:p></p>
<p class=MsoPlainText>Toronto (Visiting Assistant Professor, 1994-1997), and
the Computer <o:p></o:p></p>
<p class=MsoPlainText>Vision Laboratory of the Center for Automation Research
at the <o:p></o:p></p>
<p class=MsoPlainText>University of Maryland (Assistant Research Scientist,
1993-1994, <o:p></o:p></p>
<p class=MsoPlainText>Visiting Assistant Professor, 1994-1997). Prior to his
academic career, <o:p></o:p></p>
<p class=MsoPlainText>he worked in the computer vision industry, designing
image processing <o:p></o:p></p>
<p class=MsoPlainText>systems for Grinnell Systems Inc., San Jose, CA,
1983-1984, and optical <o:p></o:p></p>
<p class=MsoPlainText>character recognition systems for DEST, Inc., Milpitas,
CA, 1984-1985.<o:p></o:p></p>
</div>
</body>
</html>