http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
VR 환경에서 눈 주위의 임의 위치에서 측정된 다채널 안구전도 신호를 이용한 실시간 시선 추적 방법
임창환(Chang-Hwan Im),최강민(Kang-min Choi),차호승(Ho-Seung Cha),이상준(Sangjun Lee),박지민(Jimin Park),김수혜(Suhye Kim) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.8
With the development of interactive virtual reality (VR) applications, eye tracking technology has been gaining attention as one of the promising interaction interfaces in VR environments. Although most eye tracking studies in VR environment were conducted using infrared camera owing to its superior accuracy and resolution, eye trackers are generally too expensive to be manufactured as consumer-electronic devices. On the other hand, electrooculogram (EOG) has the potential to be used to track eyeball movement, with its cost and portability much better than the camera-based eye trackers. Conventional EOGbased eye trackers used two pair of electrodes to record horizontal and vertical components of EOG; however, recently developed VR head-mounted display (HMD) headsets do not always locate electrodes on the typical EOG recording sites (e.g., upper and lower part of an eye; left and right side of both eyes) because the electrodes are also used for other applications such as electromyogram (EMG)-based facial emotion detection or EMGbased silent speech recognition. In the present study, we present a new eye tracking method using 8-channel EOG signals recorded from arbitrary locations around eyes. For the comparison of the performance of eye tracking, we simultaneously recorded EOG signals from the conventional 4-channel electrode configuration (up, down, left, and right side of the eyes). To estimate the vertical and horizontal eye movement components from 8-channel EOG signals, reconstruction independent component analysis (rICA) algorithm was employed. The performance of the proposed eye-tracking system was evaluated by applying it to the real-time eyewriting system. Experimental results showed that the pattern classification accuracies obtained from the proposed and conventional methods were almost equivalent (94.1% for conventional electrode configuration and 94.4% for 8-channel electrode configuration), demonstrating that reliable eye-tracking is possible using the arbitrary multichannel EOG around the eyes.