Prototypes

Sergeant Blackwell

2004-2008
Project Leader: Diane Piepol

Sergeant John Blackwell, Interactive Character is a 3D virtual character capable of spoken interaction using ICT natural language processing technology. He appears at human scale on a transparent digital flat display system developed by ICT Mixed Reality Research and Development.

Showcasing the state-of-the-art virtual human and graphics technology developed at ICT, this work accelerated ongoing efforts to present an engaging, interactive avatar combining natural language, speech recognition with audio-driven, real time facial animation research.

The emphasis of the live demonstration is on the character’s interaction capabilities. Using ICT’s unique spoken dialogue system, the character engages with a demonstrator in conversation in a query and response format.

With his extensive vocabulary and artificial intelligence language programming, Sgt. Blackwell can answer a wide range of questions, structuring his answers based on recognized words in the questions. Engaging and funny, Blackwell even answers open-ended questions. For example, when asked, “What time is it?” he may answer, “What am I, a clock?” Or, if asked about the weather, he may reply, “You are asking me the wrong question. Why don’t you try asking me about my technology?

The USC ICT Mixed Reality Research and Development team produced this demonstration showcasing virtual human technology for the 24th Army Science Conference. He made his public debut November 28th – December 2nd, 2004 in Orlando and then traveled the country, appearing in the Presidential Classroom, an honors program that brings high school students to Washington, D.C.. Blackwell also  spoke with visitors at the Smithsonian’s prestigious Cooper-Hewitt Museum in New York as part of their National Design Triennial, Design Life Now exhibit. The exhibit was also be seen at the Institute of Contemporary Art, Boston and will traveled  to the Contemporary Arts Museum, Houston (January 2008).

Goals

  • Create a believable virtual human capable of real-time, spoken interaction on a limited content domain.
  • Present this character life size in a convincing, mixed reality display environment.

External Collaborators

  • David Hendrie
  • Dick Lindheim
  • Ernie Eastlund
  • Peter McNerney
  • Kia Zokaie