World Scientific
  • Search
  •   
Skip main navigation

Cookies Notification

We use cookies on this site to enhance your user experience. By continuing to browse the site, you consent to the use of our cookies. Learn More
×

System Upgrade on Tue, May 28th, 2024 at 2am (EDT)

Existing users will be able to log into the site and access content. However, E-commerce and registration of new users may not be available for up to 12 hours.
For online purchase, please visit us again. Contact us at [email protected] for any enquiries.

A REAL-TIME ROBOT VISION APPROACH COMBINING VISUAL SALIENCY AND UNSUPERVISED LEARNING

    https://doi.org/10.1142/9789814374286_0028Cited by:6 (Source: Crossref)
    Abstract:

    Facing the complexity of the real world, no robot can rely solely on pre-programmed knowledge and human-prepared data. If we are to create a truly autonomous robot, learning on-the-fly in the real environment is a must. The aim of this paper is to present a novel hybrid approach to unsupervised real-time learning of objects in context of mobile robotics. We develop on techniques inspired by human visual system and by research on how human infants learn. We combine our contribution on salient object detection with state-of-the-art object recognition algorithms in order to acquire both fast learning and recognition capabilities for a humanoid robot. To test it, verification on the MSRA Salient Object Database benchmark is carried out as well as several experiments with learning generic objects in a real office environment using humanoid robot Nao.