Prediction of interaction intention based on eye movement gaze feature
Bo Hu, Xiaowei Liu, Wei Wang, Rui Cai, Fangzheng Li, Shuzhi Yuan · 2019
In this paper, an interactive intention prediction method based on eye movement gaze features of typical interaction paradigm of user interface was proposed. Firstly, the common graphical user interface (GUI) was abstracted, and five typical interaction paradigms, Tab, Form, List, Tree and Filter were obtained. According to the characteristics of information organization structure of five typical interaction paradigms, five experimental task interfaces and tasks were designed, and eye movement gaze characteristics were recorded by eye tracker during the experiment. Cluster analysis was conducted on the subjects' fixation coordinates under different experimental interface conditions, and the spatial distribution characteristics of fixation were obtained. The clustering analysis results were used as the input of support vector machine (SVM), and the typical interaction paradigm types which subject were gazed are predicted. The 64% accuracy rate is obtained, which verifies the effectiveness of the proposed method.