

|
Demonstration of a very stable Poster Tracker
This Poster Tracker guesses the pose of a printed poster. This pose can be used to project 3-dimensional objects on this poster, just like in this video. Therefore any poster can be “learned” and detected with an camera. It is extremely stable, even during heavy camera movements and high angles to the poster. Depending on the poster it is very robust against occlusion, as a very small part of the image is sufficient to estimate the pose.
|
|
Continous Natural User Interface (CNUI)
Augmented reality (AR) presentation enables the creation of natural
user interfaces that employ the whole user’s environment as interaction
device. Additionally, by using hand based 3D interaction
with gestures that have a physical meaning like grabbing, dragging,
and dropping this leads to a user experience that is intuitive, since
close to the real world’s behavior. We propose a novel approach
to an AR-based natural user interface, that goes one step further
by enabling the contents of the interface to switch domains from a
virtual instance in AR to a physical instance in the real-world. All
instances stay associated and changes made to the physical instance
will be reflected on the virtual one. Because the behavior of our interface
in AR is in key aspects consistent with the real-world, the
gap between those domains is made less salient.
|
|
Fast Hand Detection using Posture Invariant Constraints
|
|
Video-based Interaction for Augmented Reality
We explored a new way to query process and memory data from real objects using Augmented Reality and Natural Interaction.
The technology is demonstrated in the context of an industrial application but many diverse applications ranging from training, education, to marketing and entertainment are of course possible.
|
|
Tracking and interaction over fisheye
The basic idea is to capture interaction and user head position/orientation over one single sensor – a camera equipped with a fisheye lens. The video shows early results which consist of the tracking over one single circular marker, and detection of the right and left hand.
|
|
Integral P-Channel for object recognition
In this video, the object's region is described using the P-Channels representation. A reference image with manually marked region is provided by the user.
In order to find the position of the screw gun in a new image of the sequence, we exhaustively search the image for the best match with the reference region. This exhautive search is made possible by the use of integral image techniques that we adapted for the P-Channel representation. Note that the video frames are processed independently (no temporal tracking).
|
|
Robust detection of the hand and pointing direction
Image based interaction represents an attractive way of interacting with the computer. However robustness is the main challenge of this kind of approach. In this video we show a very robust way to detect the hand and derive the pointing direction of the hand. Note the very similar color of the background (the table) and the hand.
|
|
Circular markers for fisheye cameras
Using Circular Markers allow for a detection directly in fisheye cameras. Once the ellipse is detected, the contour is locally wrapped into a perspective view in order to find the camera pose.
|
|
Circular Marker Tracker
This video shows our first experiments about a circular marker system.
The position of the camera can be extracted from one single marker using the properties of the ellipse extracted from the marker's contour.
|
German Research Center for Artificial Intelligence
Research Department Augmented Vision
Prof. Dr. Didier Stricker
Trippstadter Str. 122
D-67663 Kaiserslautern
Tel.: +49 (0)631 20 57 53 50
Fax: +49 (0)631 20 57 53 52
Didier [dot] Stricker [at] dfki [dot] de