We have developed an attention-focusing user interface (UI) that offers a new service for connecting real objects to digital data on the Internet. The attention-focusing UI is required to quickly and correctly recognize planar objects around users. However, conventional methods have several problems, such as the high computational cost and long processing time to extract the feature of the planar objects from images. To overcome these problems, we propose a method for extracting region-based edge features that can quickly and robustly recognize planar objects. We have conducted verification tests using smartphone and tablets and confirmed that a recognition accuracy of 99% accuracy is achieved with a response time of 0.5 seconds.
[Publications (Japanese) ]