*********
Darshak: A Planning System for Cinematic Visual Discourse Generation in Virtual Environments
Abstract (292 words)
*********
Narrative is one of the fundamental ways in which humans organize information about the world around them. 3D virtual environments -- now widely used for entertainment and education -- provide a compelling new medium for creating and sharing narratives. In pre-rendered views of virtual environments like animated movies, directors communicate complex narratives by carefully constructing the visuals shot-by-shot. Film directors exploit the user's familiarity with narrative patterns and cinematic idioms to effectively convey stories. In real-time environments like games and training simulations, camera movements used to effectively communicate a story cannot be crafted ahead of time. Because the user's actions directly contribute to the construction of the narrative in a branching story space with a possibly infinite number of story trajectories, truly dynamic systems must rely upon methods for the automatic generation of camera control.
In this talk, I will describe an intelligent camera planning system - Darshak - that takes as input a representation of a storyline and constructs visual narrative discourse for communicating that story in a 3D virtual environment. Darshak incorporates a representation of situational dramatic patterns of storytelling as abstract communicative operators. The approach uses a hierarchical partial order causal link planning algorithm to generate narrative plans that contain combined story and discourse actions. In this model, camera shots and transitions represent cinematic conventions and are defined as execution primitives. Representation of narrative discourse as a hierarchical plan structure enables us to utilize 1) the hierarchical nature of narrative patterns and film idioms through the hierarchy in decompositional plan operators, 2) the explicit representation of causal motivation for selection of shots through causal links, and 3) the explicit representation of cinematic discourse structure to make relative comparisons between candidate shot sequences when searching to construct a cinematic plan.
Bio (Long - 198 words)
***
Arnav Jhala is a PhD Candidate in the Department of Computer Science at North Carolina State University (NCSU). His research interests lie at the intersection of Artificial Intelligence and Digital Media, particularly in the areas of Computer Games, Cinematic Communication, and Narrative Discourse. In his dissertation research he has developed computational models of film idioms and algorithms for automatically generating cinematic discourse. Arnav has previously worked as an intern at the Institute for Creative Technologies at University of Southern California where he was part of the Leaders project developed in association with Paramount Pictures. He spent a summer working on the America's Army: Adaptive Thinking and Leadership game at Virtual Heroes, Inc., a leading serious games developer. Arnav is a recipient of the Outstanding Teaching Assistant award at NCSU for his work as the TA for the capstone projects class in game development and has also served as the instructor for this course. Prior to graduate school, he lived in Ahmedabad, India where he obtained a Bachelors degree in Computer Engineering with distinction from Gujarat University. He spent a year at the Indian Space Research Organization (ISRO) in the Space Applications Center working on his undergraduate senior project.
Bio (Short - 62 words)
Arnav Jhala is a PhD Candidate in the Department of Computer Science at North Carolina State University. His research interests lie at the intersection of Artificial Intelligence and Digital Media, particularly in the areas of Computer Games, Cinematic Communication, and Narrative Discourse. In his dissertation research he has developed computational models of film idioms and algorithms for
automatically generating cinematic discourse.