next up previous contents New Search Help FLUIDS_Home DFKI_Home
Next: Knowledge-based Graphics Generation - Related Up: Knowledge-based Graphics Generation Previous: Knowledge-based Graphics Generation - Research


Knowledge-based Graphics Generation - Research and Market Trends in Japan and USA

  Research efforts in Japan concentrate mainly on 3-dimensional computer graphics. In particular, the ongoing investigations concerning the generation and manipulation of 3-dimensional displays from natural language input are noticeable. Current activities in the commercial sector point to an interesting market potential for shared virtual environments, where a knowledge-based approach to graphics generation will be essential.

Separator_Line

In the USA, both 2D as well as 3D graphics generation is addressed in current research. One focus is on the automatic and semi-automatic generation of business graphics. Obviously, these efforts target at the important market for business intelligence systems. Apparently, knowledge-based methods for 3D animation constitute another focus for research and development. Improved capabilities in these area can be regarded as a decisive requirement for the next generation of multimedia interfaces and simulation systems.

Separator_Line

(New Sep. 96) A recent issue of ACM Communications , the flagship publication of the Association for Computing Machinery , makes clear that in the USA intelligent graphics is now seen as one of the new important paradigms for computing. The contribution by Lieberman provides some convincing examples of real applications and advocates the amplification of current research and development activities in the area of knowledge-based graphics generation.

Separator_Line

(New June 97) Given the US dominance in the field of knowledge-based graphics generation it is not surprising that there are also excellent research groups working on virtual humans. As in Europe, a current focus of research is on intelligent autonomous and interactive artificial humans. In the USA, however, the market for virtual humans seems already much better developed. In particular, the entertainement industry seems to play an important role for accelerating the commercial exploitation of this new technology.

As the example of the world's first virtual actress and singer Kyoko Date shows, virtual humans will also influence the entertainement industry in Japan. Kyoko Date , the most photorealistic virtual human to date, is a product of Japan's top modeling agency Hori Productions Inc. and has gained an impressive market success in selling music videos and a CD. The project was initiated in 1996 and is continued so that in the future emerging technology will enable the `virtual idol' to appear on a live TV show and chat with other artists.

Separator_Line

(New July 98) AVI-98, which constitutes one of the leading conferences specifically devoted to advanced visual user interfaces, attracted many contributions from the USA. Compared with European research activities presented at the Advanced Visual Interfaces Conference current research in this field in the USA seems to focus more on themes like navigation in information spaces and multimedia design. However, similar to the trend in Europe it can be observed that the general field of advanced visual user interfaces is strongly influenced through Artifical Intelligence approaches like knowledge-based graphics generation and intelligent interfaces.

Separator_Line

(New Nov. 98) In the USA, the growing convergence of graphics and vision is regarded as a salient trend in current research and development which will also influence knowledge-based graphics generation in particular.
Computer graphics and computer vision are inverse problems and in the traditional view computer graphics starts with input geometric models and produces image sequences whereas computer vision starts with input image sequences and produces geometric models. However, as the article Lengyel [1998] illustrates vision researchers now work from images backward just as far as necessary to create models that capture a scene without going to full geometric models. In order to create stunning, photorealistic images in real time graphics generation works more and more with hybrid geometry and image models. These models use images as partial results, reusing them to take advantage of similarities in the image stream. Lengyel, who works for Microsoft Research , mentions several reasons for the convergence of graphics and vision. Combined graphics and vision approaches have a hybrid vigor, much of which stems from sampled representations. The use of captured scenes yields richer rendering and modeling methods than methods that synthesize everything from scratch. Exploiting temporal and spatial coherence, i.e. similarities in images, via the use of layers and other techniques is boosting runtime performance and finally the explosion in PC graphics performance is making powerful computational techniques more practical.


next up previous contents New Search Help FLUIDS_Home DFKI_Home
Next: Knowledge-based Graphics Generation - Related Up: Knowledge-based Graphics Generation Previous: Knowledge-based Graphics Generation - Research


Gerd Herzog
Last update: Fri Feb 26 13:17:30 MET 1999


Send comments to herzog@acm.org