Describing 2D Objects by using Qualitative Models of Color and Shape at a Fine Level of Granularity

Zoe Falomir, Jon Almazan, Lledo Museros, M. Teresa Escrig

Service robots need a cognitive vision system in order to interact with people. Human beings usually use their language to describe their environment and, as qualitative descriptions can be easily translated into language, they are more understandable to people. The main aim of this paper is to define an approach which can obtain a unique and complete qualitative description of any two-dimensional object appearing in a digital image. In order to achieve this, first, Museros and Escrig’s approach for shape description is extended, secondly, a characterization of the objects in the image according to its regularity, its convexity and the number of edges and kind of angles that its shape has, is explained, and finally, a qualitative model for color naming based on HSV coordinates is defined. An application that provides the qualitative description of all two-dimensional objects contained in a digital image has been implemented and promising results are obtained.

Subjects: 3.5 Qualitative Reasoning; 19. Vision

Submitted: Apr 30, 2008


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.