The technique of Projection Mapping, which is useful for merging real-world geometry with an augmented appearance, is a promising core technology for augmented reality (AR). In recent years, dynamically changing environments, mainly a consequence of the growing demand for interactive user experiences, have contributed to a new style of AR applications. However, performance levels of current systems for realizing 3D effects, in terms of the tracking speed and projection ability, are insufficient to meet these demands. In this paper, we present a high-speed, occlusion-robust marker-based 3D tracking method achieved by only using a monocular monochrome image. The objective of our research is to develop an automatic marker design method for any 3D shape and an effective framework for stabilizing tracking at high throughput by extending the latest promising work based on a deformable dot cluster marker [46]. Furthermore, this tracking method was used in combination with a high-speed projector, both of which can achieve high throughput and low latency, on the order of milliseconds. This enabled us to realize a high-quality computational display capable of representing the material appearance of a dynamically moving target. The demonstration showed that the effect of a dynamically changing appearance with nearly imperceptible latency drastically enriches the sense of immersion in the recognition of augmented materials with the naked eye.

We developed an application specific multi-processor generation system intended for realtime applications. In this system, we adopted a distributed memory type multi-processor architecture with hierarchical tree network as a configurable multiprocessor which can be adapted to various scale systems flexibly. We have also developed a configurable multi-processor prototype as LSI chips with the 0.18 /spl mu/m CMOS standard cell technology.