To provide a technology for prospecting a parameter for a gazing point generation model from time series of gazing point measured by an eye tracker.SOLUTION: A true gazing point by a person at a time point t is defined as follows. In a case where a saccade occurs at a time point near the time point t, a point transferred from a true gazing point at a time point t-1 in a direction and at a length depending on the saccade is defined as a temporary position. In a case where no saccade occurs at a time point near the time point t, a point transferred from the true gazing point at the time point t-1 in a random direction and at a random length is defined as the temporary position. The true gazing point by a person at the time point t is defined as a point which is obtained by correction of the temporary position with a gaze delay at the time point t, which follows AR (2) model. A model parameter for a gazing point generation model is prospected from time series of gazing point obtained by measuring eye movements of a subject, in which the gazing point generation model is modelized by defining an actually measured gazing point at the time point t as that obtained by adding noise to the true gazing point at the time point t.SELECTED DRAWING: Figure 5【課題】アイトラッカーによって計測された注視点の時系列から注視点の生成モデルのパラメータを推定する技術を提供する。【解決手段】時刻ｔの人の真の注目点は、時刻ｔ近傍にサッカードが発生する場合は時刻t-1の真の注目点を当該サッカードの方向及び大きさに応じて移動させた位置とし、時刻ｔ近傍にサッカードが発生しない場合は時刻t-1の真の注目点をランダムな方向及び大きさに応じて移動させた位置とし、時刻ｔの人の注視点の遅れは、AR(2)モデルに従うものとし、時刻ｔの人の真の注視点は、時刻ｔの真の注目点を時刻ｔの注視点の遅れにより補正した位置とし、実際に計測される時刻ｔの注視点を、上記時刻ｔの真の注視点にノイズが加わったものとしてモデル化したものを注視点の生成モデルとして、対象者の眼の動きを計測して得た注視点の時系列から、注視点の生成モデルのモデルパラメータを推定する。【選択図】図５
