Research On Video Switching Algorithm Based On Eye Movement Information Of Single Object

With the advancement of the times,science and technology are becoming more and more developed,and the way of human-computer interaction is constantly developing and improving.80% of the external information we obtain comes from eyes,and using eyes to achieve human-computer interaction is the easiest and fastest way.With the emergence of short video apps,more and more people spend time on short videos,among which there are many elderly and disabled people with inflexible fingers,and it is a little difficult to switch videos.Therefore,it is necessary to study the video switching algorithm based on eye tracking information,which can be used to determine the user’s preference for the video,so as to recommend the video for the user according to the preference,switch the video for the disabled,promote video advertising,and judge the user’s attention to the product by judging the eye movement information of the user watching the advertisement.At this stage,the algorithms that implement these functions have certain limitations,such as the need to use expensive instruments such as eye tracker equipment,the need for users to provide additional actions,and the use of infrared light source assistance.In this paper,an active video switching algorithm based on eye tracking information is implemented without additional light source and only inexpensive camera.The research content of this paper is as follows:1.This paper uses the face detection algorithm based on Adaboost to realize face detection and dlib face landmark point positioning to obtain accurate face landmark point coordinates.The position coordinates of the pupil are extracted by combining gradient information and random circle.Then the position coordinates of the pupil are obtained by clustering the center of the random circle by clustering the center of the random circle in the part based on the random circle.Finally,a robust and accurate iris center is obtained.2.In the video active switching algorithm,this paper uses a coarse judgment method based on eye posture.It reduces the time for precise judgment,so as to realize the time series alignment of the video and eye tracking information extraction part and ensure the real-time performance of the algorithm.The POSIT-based algorithm is used to estimate the head pose,reduce the influence of head pose on the line of sight,obtain the translation matrix and rotation matrix of the initial position of the head and the current position,and find the pitch and yaw deflection angles of the head.3.In this paper,an active switching algorithm is proposed to judge the ROI region and screen visibility angle of the human eye using spatial geometry,and complete the video operation by judging the result.The experimental results show that the interactive system can make corresponding operations on the video according to the change of human gaze,and realize convenient and concise human-computer interaction.