Prediction of user’s intention based on the eye-hand coordination

The action to grasp an object by a human hand comprises not only an eye-hand coordination, but also a motion process that controls and plans the movement for grasping objects. To understand this behavior, it is necessary to study eye and hand movements simultaneously. A novel approach to understand the user’s intentions with computer vision technics is proposed in this paper. Our solution performs a fusion between the user’s gaze and a hand motion estimation using wearable systems and multiples cameras. To realize such system, we employ multiples cameras on the user’s head, beneath the wrist, and a eye-tracker device. To predict the user’s movement, we use information from multiples point-of-view and we characterize a handmovement trough a Hidden-Markov Model framework. As main result, our system can predict the movement before the hand can reach an