In the recent years, timely and accurate assessment of classroom teaching has become an interesting research area. Its focus is to promote quality of classroom teaching by enhancing instructor’s lecture delivery skills by providing timely feedback and suggestive measures. Involvement/engagement of students in the lecture contents and students’ understanding of the lecture material are used to measure the effectiveness of the teaching methodology. Both of these measures can be correlated with students’ facial expressions.

Usually students have some sort of smiling gesture on their faces when they understand the lecture content and in case of complex contents, students shows negative expressions such as disgust, sadness etc. Facial expressions serve as a mode of nonverbal communication and an instructor can adjust his/her teaching methodology by comprehending facial expressions of the students.

This research project is aimed to automatically evaluate the students’ comprehension level in the real time classroom environment. For this purpose, three HD CCTV cameras will be installed in a classroom to obtain facial images of all the students from different angles. Face image from the live video streams for each student will be detected and facial features will be extracted accordingly.

These facial features will subsequently be used to determine seven different types of facial expressions using machine learning technique. Besides these facial expressions, face action units will also be extracted. Finally, face action units and facial expressions will be fused together to make a decision about the student comprehension level of the lecture contents. This assessment then be communicated to the instructor and management to take corrective actions in case of low comprehension level of students is identified.