G-2-3 カメラ画像に基づく虹彩の自動検出法(メカニカルシステムとその知能化(2),口頭発表)

抄録

This study aspires realization of eye-gaze tracking to provide human instructions to robot through eye-gaze In recent years,although the eye-gaze detection study becomes active,its accuracy doesn't enough to provide instructions to robot.Furthermore,since detection of eye is easily affected by light fluctuation,there are many special devices capable of measuring the eye-gaze with high accuracy.Hence,this study's objective is derivation of new algorithm.which doesn't depend on external device.To realize this objective,we detect the eye-gaze using general-use Web- camera to obtain image and perform image processing using OpenCV Especially,we produce a new iris-detection method.which is used for the eye-gaze detection as a key technology.Since in 40 trials average brightness of true inner circle is less than one of false iris inner circles in probability of 97.5%,this new algorithm is available for the eye-gaze detection.