Poster: Say it to see it: A speech based immersive model retrieval system

Abstract

A B C Figure 1: Demonstration: The user speaks the term "pizza" and is presented with a visual interface of results (A). The user selects an item and a placeholder box is created that the user can interact with (B). Upon model load, the placeholder box is replaced with the model (C). ABSTRACTVirtual spaces have proven to be a valuable means to visualize and inspect 3D environments. Unfortunately, adding objects to 3D scenes while inside of an immersive environment is often difficult as the means used to acquire models from repositories are built for standard computer interfaces and are generally not available during a users session. We develop a novel interface for the insertion of models into a virtual scene through the use of voice, 3D visuals, and a 3D input device. Our interface seamlessly communicates with external model repositories (Trimble 3D Warehouse) enabling models to be acquired and inserted in the scene during a user's virtual session. We see the benefits of our pipeline and interface in the fields of design, architecture, and simulation.