ヘッドマウントディスプレイと学習理論に基づく画像認識によるインタフェースシステム

抄録

In this paper, we have developed a new human interface device using a raspberry pi (single board computer) and a head mounted display aiming for use in physically handicapped persons, factories, medical sites, etc. This system consists of raspberry pi, head mounted display and a CMOS image sensor camera, and recognizes the shape of the hand gesture using a cascade classifier which is one of machine learning algorithms using OpenCV which is an image processing library. According to the recognized hand gesture, various image manipulations which are switching of browsing images / enlargement / reduction / movement / rotation etc., with respect to design drawings and surgical photographs saved in advance, and photographing of images acquired with the CMOS image sensor camera, can be performed. Verification experiment using this system confirmed that this system works effectively. By using a hand gesture, it became possible to perform various image manipulations during work such as a handicapped person, a factory, a medical scene or the like moreover irrespective of age and sex.