EXPRESSIO: AUTOMATIC FEEDBACK FOR MOOC TEACHERS BASED ON AFFECTIVE COMPUTING
EXPRESSIO: AUTOMATIC FEEDBACK FOR MOOC TEACHERS BASED ON AFFECTIVE COMPUTING

Summary/Abstract: The increase in popularity of Massive Open Online Courses (MOOCs) raises the issue of effective student feedback. Instructors can examine students’ opinions and evaluations through multiple methods, such as eliciting them in feedback forms, or analyzing statistics that are automatically generated on the course platform about students’ participation and achievements. While feedback forms can provide valuable information on students’ learning experiences, including their interest and emotional reactions, they are vulnerable to low response rates and to various forms of recollection bias. For example, human memory privileges the final elements of an activity over its initial or middle periods (Kahneman 2010). Eliciting feedback on an entire activity also lacks the granularity required to determine what specific aspects were most and least interesting. While the analysis of automatically generated indicators compensates for the low response problem, it also fails to deliver finely-tuned information on student engagement with specific elements of the course. In order to address this issue, we developed Expressio, an automatic student feedback solution based on affective computing. We start from Ekman’s emotion classification, distinguishing six basic emotions: happiness, sadness, surprise, fear, anger and disgust. We rely on several technologies and devices: Creative Senz3D Camera, Intel Perceptual Computing SDK, OpenCV, Windows API and Microsoft Visual Studio 2012. Expressio is a program that identifies user expressions as they occur during an online activity and displays them on the screen in a GUI. Our solution also affords training the program for a more accurate and personalized identification of emotions. Expressio can be integrated with a MOOC to allow the transmission of continuous feedback regarding student interest and emotional experiences during the course, based on automatic detection of facial emotional expressions.

• Details
• Contents