Overview: Sensing, 3D reconstruction, virtual reality, vision classification

    - Using the Kinect/Xtion(K||X) to find the hand using the NITE module. Then to follow this by centering it in the image by sending serial commands to the turret. Can this be used for interaction

    - Using the K||X to find the face using openCV,, then centering this in the image by sending serial commands to the turret

    - To move the turret with the K||X on the end and build up a reconstruction of the surrounds. Using EyeInHand or PCL to fuse the points.

    - Find objects in the 3D data using combination of 3D and RGB data. And algorithms to find the object based upon matching templates, shapes, contours, SIFT features.

    - To find grasping parameters of the objects in the images and perform a manipulation task

    - To map the surrounds in 3D for use in collision avoidance (i.e. and occupancy map of the surrounds)

    *- To move the head in such a way as to maximise the information found and to keep the information as up to date as possible (mapping, coverage and exploration)

    - Specify areas in the world that are more important so as to keep those voxels more unto date that others.