Visual comfort improvement for 3D video based on parallax adjustment

This paper presents a parallax adjustment method for 3D video to improve the visual comfort of the 3D video. Firstly, matching algorithm based on SIFT (Scale Invariant Feature Transform) is used to obtain the accurate horizontal parallax and vertical parallax of the feature points between the left-view and the right-view of the 3D video. Then the range of horizontal parallax with comfort viewing is determined according to statistics of the horizontal parallax of the whole video. Finally, two 2D projective transformations are used to achieve vertical parallax elimination and horizontal parallax adjustment respectively. Experimental results show that the average subjective evaluation score of the improved 3D video increases about 4.82%.