A Tabletop Interface Using Fingernail Images and Real Object Recognition

摘要

A variety of research on tabletop interfaces has been reported in the last decade. In a tabletop system, the difference in operation between digital media and physical media such as paper gives an unsatisfactory experience and disturbs their comfortable simultaneous use. For users to be satisfied, they must be able to operate the digital and physical media seamlessly. In this paper, we propose a tabletop interface system that allows users to operate digital media intuitively by using gesture recognition. The proposed table can capture an image of an object on the table by utilizing press gesture recognition from images of a user's fingernail based on support vector machine (SVM), and can acquire digital content from digital paper by utilizing a shake gesture of the device. The evaluation experiments show that the proposed system recognizes a user's finger state with high accuracy and that the users can intuitively operate the table without any awareness of data transmission.