The Augmented Painting

An interactive tabletop augmented-reality setup that creates an artistic representation of the scene. Users can select a painterly, cartoon-like, or black-and-white illustrative style. Virtual objects are added and displayed in the same artistic style.

Enhanced Life
Augmented Painting combines augmented-reality techniques with real-time artistic stylization methods. This concept ("stylized AR") is still very new and was only recently introduced. The overall setup integrates video acquisition, video-based camera tracking, artistic video stylization, and artistic rendering in a single interactive, real-time system.

To create a novel augmented environment. Because real and virtual scene elements are equalized, they are difficult to distinguish. This leads to a new perception of the augmented scene and a new experience.

Several new stylization algorithms were developed in order to generate a stylized augmented video stream at real-time frame rates. Some of these algorithms utilize the programmability of modern graphics processing units, which generates a high-speed, high-quality video stream that would have been impossible only a few years ago.

Several other interesting potential applications can be conceived for stylized augmented environments. Combined stylized representations of virtual and real objects can provide a novel experience in augmented-reality games, entertainment scenarios, and education and training applications. Since physical and graphical objects are difficult to distinguish in this type of augmented video stream, this might also lead to new impulses for immersive AR settings using head-mounted displays or translucent display setups. Since the user cannot tell whether an object in the immersive augmented environment is real or virtual, the technology might actually become more effective for some applications thanks to the stylization. We are currently considering potential uses in psychotherapy (for example, phobia therapy) and psychophysical studies (for example, face recognition in virtual environments).

Jan Fischer
WSI/GRIS, Universität Tübingen
fischer (at)

Dirk Bartz
Wolfgang Strasser
Universität Tübingen