<html>
<head>
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<font face="Palatino">Wednesday, March 13, Noon, ASH Lobby<br>
<br>
William Bares, Ph.D.<br>
Candidate for Assistant Professor of Computer Science and Game
Design<br>
<br>
Talk Title: <b>Fear not the zombies! Or, Cinematography by
Human-Computer Dialog</b><br>
<br>
Abstract: What if a movie camera could learn about cinematography
while an expert human filmmaker operates the camera? Could that
same movie camera, now in the hands of a novice, offer the novice
filmmaker suggestions on composition and editing in the style
learned from the expert?<br>
This colloquium explores these questions by presenting a novel
cinematography assistant embodied in a portable touch screen
equipped with motion-sensors that is used to move and aim a camera
in a computer-generated world. As a filmmaker lines up shots by
turning and moving the portable screen as if it were a camera
viewfinder, the assistant automatically analyzes shot properties
and transitions between shots to formulate a continually evolving
model of the filmmaker’s editing style. In preparing to film the
next shot, the filmmaker can request suggestions presented as a
palette of distinct shot angles and distances from which to begin
the next shot. The assistant computes suggested shots so that
they reflect established practices of continuity such as not
“crossing the line” and matching eye looks of interacting
characters. Suggested shots also reflect the user’s past editing
tendencies when cutting from one shot to the next. The filmmaker
selects a suggestion by touching its image in the palette of
suggestions and is instantly transported into the suggested shot.
Then, the filmmaker can either accept the suggested shot as-is,
ignore it to compose his or her own, or modify it by moving or
turning the portable touch-screen. <br>
Future directions of work invite participation by anyone having an
interest in film, animation, artificial intelligence, and
interfaces for games, new media, graphical simulations, or
learning environments. <br>
<br>
Biographical Statement: William Bares enjoys making connections
between the arts and technology in teaching, research, and
community engagement. <br>
He envisions future games and interactive media experiences that
take place in real-world spaces in which human players and avatars
interact through speech, motion, and gestures. His current
research seeks to develop methods to augment the creativity and
productivity of cinematographers. He earned his Ph.D. at North
Carolina State University while studying with the
interdisciplinary IntelliMedia Initiative which creates
intelligent virtual 3d learning environments. He has been an
invited researcher at INRIA (Rennes, France) and most recently at
the Games and Playable Media Group at UC Santa Cruz. He is
currently a faculty member in Computer Science at Millsaps
College.</font><br>
<div class="moz-signature">-- <br>
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
<title>CBD to-do list updated and attached</title>
Paula Harmon, Administrative Assistant <br>
<div class="moz-signature"><small> School of Cognitive Science <br>
Hampshire College<br>
893 West Street Amherst, MA 01002 <br>
phone: 413.559.5502 <br>
fax: 413.559.5438 <br>
<a href="http://cs.hampshire.edu">http://cs.hampshire.edu</a></small>
</div>
</div>
</body>
</html>