User attention analysis for e-learning systems using gaze and speech information

In this paper, a practical approach to detecting user attention for online learning systems using gaze and speech activity information is proposed. A portable camera and microphone system to acquire human-computer interaction data is used in this approach. User attentiveness is classified into three categories: attentive, inattentive-not looking, and inattentive-speaking. Experimental results show that the proposed system clearly differentiates the attention level of users that participate in e-learning sessions. The proposed system is suitable for realtime applications.