Research of Rapid Attention-Information Extraction in Human-Robot Interaction

In the human-robot interactions, it is an important research on the artificial intelligence that robots catch the human´s attention information. This process has a significant impact on robot intelligence. In this paper, a system for extracting human´s attention-information from interaction has been designed. On the basis of face detection and eye-area location, an algorithm, combined with ant colony algorithm(ACA) eye edge extracting, curve fitting and eye-model 3D reconstruction, has been designed to achieve attention information successfully by using an ordinary web cam and a low requirement hardware system. There is a significant help on real-time and applicability for human-robot visual interaction.