Reality Tooning: Fast Non-Photorealism for Augmented Video Streams (poster

Recently, we have proposed a novel approach to generating augmented video streams. With this new method, the output images are a non-photorealistic reproduction of the augmented environment. Special stylization methods are applied to both the background camera image and the virtual objects. This way, the graphical foreground and the real background images are rendered in a similar style, so that they are less distinguishable from each other. Here, we present a new algorithm for the cartoon-like stylization of augmented reality images, which uses a novel postprocessing filter for cartoon-like color segmentation and high-contrast silhouettes. In order to make a fast postprocessing of rendered images possible, the programmability of modern graphics hardware is exploited. The system is capable of generating a stylized augmented video stream of high visual quality at real-time frame rates. 1. Stylized Augmented Reality In conventional augmented reality systems, the graphical objects are rendered over the camera image using standard real-time graphics algorithms. The resulting renderings can look artificial, and they tend to stand out from the background image (e.g., see Fig. 2(a)). We have recently presented an alternative approach to generating augmented video streams. Our stylized augmented reality technique attempts to create similar levels of realism in both the camera image and the graphical objects [1]. The cartoon-like stylization mode, which we have described, produces augmented images composed of mostly uniformly colored regions, which are enclosed with black silhouette lines. Here, we introduce a new algorithm for the cartoon-like stylization of AR images. This new method is a specialized postprocessing filter, which is applied to the augmented image after the overlay of virtual objects. The implementation of the new algorithm uses vertex and fragment shaders, which run on the programmable GPUs of recent graphics cards. This allows for a straightforward and efficient design of the algorithm. The new method produces a stylized augmented video stream of a better visual quality at a significantly higher frame rate compared to the previously described approach. In particular, it all but eliminates flick