PHD-IT-L Archives

October 2013

PHD-IT-L@LISTSERV.GMU.EDU

Options: Use Monospaced Font
Show HTML Part by Default
Condense Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
MIME-Version:
1.0
Sender:
Subject:
From:
Damon Mccoy <[log in to unmask]>
Date:
Mon, 28 Oct 2013 18:49:24 +0000
Content-Type:
multipart/alternative; boundary="_000_5be8af85cbe3498da0c3f98fee743f94DM2PR05MB416namprd05pro_"
Comments:
Reply-To:
Damon Mccoy <[log in to unmask]>
Parts/Attachments:
text/plain (1493 bytes) , text/html (2044 bytes)
Hello everyone,

Tim has volunteered to do a Halloween themed lunch hour seminar. I will supply Halloween themed snacks and feel free to show up in costume if you want.

Title: "Understanding the Big Picture: Aggregate Clustering for Multiple Virtual Senses"

Abstract:
Virtual characters are used to create more dynamic and lifelike virtual scenarios. These characters exist in realistic worlds consisting of a large number of objects and events, and must be able to understand their environment so that they are able to act plausibly within it. Endowing an agent with perceptual capabilities from multiple senses leads to more believable reactions to their environment.However, in complex environments with a large number of objects this can become computationally infeasible. To allow virtual humans to operate in more realistic environments, we have created two methods to cluster objects during agent perception, one based solely on the layout of objects in the environments and another based on agents' perspectives of the environment given their current mental state. Objects are grouped based on their similarities and proximity. Agents can then reason about groups of concepts,and are able to perceive a more complete picture of a given scenario.We provide sample scenarios to illustrate the impact of multisense perception on agent understanding and behavior. We also perform analysis on the impact of clustering on computation costs, using multiple methods of clustering.


ATOM RSS1 RSS2