Go to the homepage of the Vrije Universiteit. Go to the homepage of the department of computer science. Go to the homepage of the faculty of sciences.


Color-Based Object Recognition by a Grid-Connected Robot Dog

Object Learning through the Aibo Demonstrator GUI (screen shots)

This is the Aibo Object Recognition demonstrator GUI, in which the 'learning' phase is performed for one object. This is followed by a 'recognition' phase to see whether our robot dog indeed has learned the object.
The interface shows that we are using five clusters located in (1) The Netherlands, (2) Poland, (3) Austria, (4) Italy, and (5) Australia. With these clusters we obtain a framerate of around 2.9 per second. Note that we can easily increase this framerate by using more nodes per cluster, or by adding more clusters located anywhere in the world.

  • Step 1: Select the text box for inserting the name of the new object to be learned:

  • Step 2: Type in the name of the object, in this case 'hedgehog':

  • Step 3: Learn the object, by storing a reduced object representation in a local database. Our aibo robot dog will say this.

  • Step 4: Check to see whether the object indeed is being recognized. Our aibo robot dog will say this.

  • If you have any comments, please e-mail the maintainer of this page.
    Your browser does not fully support CSS. This may result in visual artifacts.