Engagement Index for Classroom Lecture using Computer Vision

Effectiveness of any interaction is determined on the basis of interaction level and understanding developed between the interacting parties. In this paper, communication between two parties: teacher and students in a classroom lecture is being considered. This classroom session can be considered as highly interactive or less interactive by testing the communication skills of the teacher, how he conveys his ideas to the students, and how well the students interpret his ideas. The main objective of the project is to find out the engagement index of teacher & student interaction in a classroom lecture. This is achieved by computer vision and machine learning techniques. Analysis of the dataset is divided into verbal and non-verbal sections, namely, audio and video recordings. The videos are used to do analysis of the students. The audios are used to analysis of teachers. Based on certain parameters like attentiveness of students during lectures, the number of questions asked by both, the students' and the teachers', the engagement index is calculated.