P2. Lifelong Fusion of Semantic Localization and Conversational skills for Social Robots in Smart Homes
Principal investigator: Ismael GarcĂa Varea
► Implementation of a scene segmentation system able to recognize and register relevant 3D objects using RGB-D perceptions to generate consistent and useful models for the CORTEX cognitive architecture.
► Development of a semantic localization system capable of describing an input RGB-D image using semantic labels that identify the location of the robot and their surrounding objects.
► Design, development and implementation of a multi-modal conversational system for social robots.
► Integration of the semantic localization and the conversational systems into a mutual process of lifelong learning.