| KEYWORD |
GR-09 - GRAphics and INtelligent Systems - GRAINS
Automatic generation of character animations and their description
Reference persons FABRIZIO LAMBERTI, ANDREA SANNA
Research Groups GR-09 - GRAphics and INtelligent Systems - GRAINS
Description This proposal can be tackled through more than one thesis works. In fact, the goal is manyfold. On the one side, the idea is to design and develop a system capable to read and understand the description of a scene in a book, a movie script, etc. (in terms of objects and characters described, their behavior, etc.) and automatically create a 3D virtual representation for it. The system shall be able, for instance, to select elements to be added to the scene from an object library, understand relationships among them (like in, on, over, etc.), set actions and attributes for them (e.g., a jumping action and a happy mood for a text like "the dog was happily hopping in the park"), etc. On the other side, the idea is to develop a tool for transforming the body movements and facial expressions of an actor into text (thus creating the book, the script, etc.). The tool will have to leverage motion capture technologies to gather information about actor's performance and exploit suitable recognition algorithms to map human activity into corresponding descriptions. Automatic text generation will rely on previous knowledge of the scene. For instance, for a scene in which the character is in front or a crowd and the actor is moving towards them in the motion capture space by smiling and swinging his hand, text could be something like "and he happily said hello to everybody there".
See also http://grains.polito.it/work.php
Deadline 13/02/2019
PROPONI LA TUA CANDIDATURA