My previous work includes methods for mapping player behaviors to game-state (here); for developing user-friendly crowd authoring tools (here); and for studying how character and avatar animation can subtly influence players (specifically, concerning naturalness, responsiveness, gaze, and emotions). I am currently studying stylistic motions for character animation at the SIG Center for Computer Graphics with Norm Badler, Sophie Jörg, and Steve Lane. The following video shows some examples I've built over the course of this research.
My resume is available here.
The reel to the left highlights some tool demos from various side projects and papers, specifically, my image painter, puff modeler and volumetric renderer, image-based tree modeling plugin for Maya, and my crowd brush demo from my 2014 I3D paper on macroscopic crowd modeling.
Bayesian Clustering of Player Styles for Multiplayer Games.
Aline Normoyle and Shane T. Jensen, Artificial Intelligence and Interactive Digital Entertainment (AIIDE), 2015
This spring, Ben Sunshine-Hill and I co-presented "How to use machine learning like a responsible adult" at the Game Developer Conference AI Summit in San Francisco. The tutorial aimed to introduce ML to people who've never looked at it before, and I hope people found it inspiring, interesting and educational!
We won best paper at Motion in
Games last fall in Los Angelos! This paper describes several
experiments we designed to better understand the trade-offs between
responsiveness and character animation in games. Two of our most interesting results were that
players felt most satisfied with our simplest animation controller and that players did not notice the
more naturalistic motions in our highest quality controller, although viewers in a separate survey did.
For more details and
results, see the paper
We have a SAP 2014 paper! This work summarizes our latest experiments investigating how delays in character responsiveness affects players' perception of themselves and the game. In this work, we look at the effects of fixed versus varying delays (e.g. jitter). We find that jitter is noticed by players more than a fixed delay and reduced the perceived motion quality of the character. Both latency and jitter hindered the player's performance, but in different ways. Check out the paper and video for details!
I presented "Stochastic
activity authoring with direct user control" at I3D
in San Francisco this week (March 14th). This technique lets
users easily author background crowd behaviors using a crowd brush. The
video below shows a demo.
An interactive version of the bird demo from the paper is also here.