This package provides the portrait_bot stack project with a GUI and the functionality to recognize and extract faces.

How does it work?

The face_detector package is used to detect faces in the wide_stereo camera's stream images. Then the tilting laser scanner is adjusted to scan the area around the face and a portrait is extracted from the prosilica camera's high-definition stream. Then a mask image is generated from the laser's point cloud. At last, both the portrait image and the mask image are sent to the face_contour_detector.

The GUI enables the user to take a portrait image of a face, let the PR-2 robot take a pen and then paint an outlined version of the portrait onto a canvas. It mainly provides information about the project's stages and allows for easy interaction.


For information on how to use the portrait_bot project, read the "Tutorial" section of the portrait_bot page.

Wiki: face_finder (last edited 2011-05-02 13:57:42 by NikolausMayer)