A Novel Video Segmentation System Based on Viewpoints

With the development of eye tracking technologies, a large number of videos with viewpoints are available. Segmenting these videos is conductive to analyzing the video content that are focused on. In this paper, we propose a video segmentation system that can extract segments which are of interest to participants based on the viewpoints recorded by eye trackers. Database is established to manage these segments for the further analysis. Experimental results show that our system can detect the change of viewpoints' types accurately and separate them effectively. In addition, we apply the system to analyze how product placements affect participants and get potential information that can only be obtained from questionnaires by traditional methods.