Face Recognition in Education

Ashu Bhatnagar October 12, 2017
Apple announced Face ID. A face recognition based authentication. There have been other players who have attempted biometric authentication in recent times (Samsung, Google, etc.). Facial recognition is finding increasing adoption as it becomes more robust. Some examples of large scale upcoming adoption of face technology are in the news (Dubai & Hong Kong Airports). In comparison with other biometric mechanisms, face recognition brings more than just identity management for education. As technology advances, new possibilities are likely to emerge. So what possibilities does this technology brings to education? In education, face recognition touches the following key aspects for learners:
  1. Privacy and security policies in education that have seen increasing focus and awareness.
  2. Personalization in education that demands identity management & usage data for analytics.
As face recognition graduates to more robust and higher maturity, what does it mean for educators, students and parents?
Technology of Facial Recognition
Face recognition is not new. But there is a likely revolution on the horizon. Reliable face recognition is trending towards becoming robust mainstream technology. Android has had face recognition for some time now. However, it has not been considered a secure solution so far. Samsung recently launched its face and iris scanning capability in Galaxy-S8. The technology for iris scanning relies on infrared light emitted by a hardware component built into the phone. A special camera takes an image of the iris flooded by this “near infrared” radiation to scan the iris. In the Samsung solution, Iris scanning is considered more secure than its face scanning technology. Apple launched face recognition via Face ID in iPhone-X recently that replaced Touch ID for authentication. Apple believes that the technology is reliable enough (even though there are doubters out there). With Apple launching Face ID, others are likely to follow soon. Face recognition as developed by Apple uses a 3D map of the face. Its use of specialized hardware for creation of 3D face scans is going to be faster and more reliable than traditional 2D imaging. A specialized hardware + software based solution is likely to be better than using traditional cameras for recognition.
Use Cases of Face Recognition
Face recognition can serve well in two different scenarios in education.  
  1. Authentication: serving like any other biometric identification mechanism but with one difference. It is better than fingerprints in that it is non intrusive when used for monitoring identity.
  2. Capture face expressions: This is an advanced scenario and has a bigger potential to help make online education better.
Let’s consider these use cases in some detail.
Face as Biometric ID – Assessments and Personalization
There are scenarios where it helps to identify the learner non intrusively. One such scenario is taking online assessments. A monitoring mechanism that polls authentication during an online test can make online assessments viable. Another use case for this scenario is to enable serving content that is personalized without need for explicit login. With recognition and identity, capturing learner usage data for analytics and personalization becomes possible using face as a biometric ID.
Learner Engagement using Face Recognition and AI
Understanding learner engagement with subject has been a difficult problem to solve. There have been experiments conducted on tracking facial expressions to capture emotions and reactions during study. In a classroom setting, this was automatically happening with the instructors engaging with learners and gauging connect in real time. For online learning, use of camera to capture facial expressions and use of AI for sentiment analysis can help determine learner’s engagement level. This is specifically useful for gauging learning effectiveness. With this technology advancing towards a more reliable face recognition, the potential to exploit this use case will increase. Apple has been investing in face technology (read here and here) for some time now. Others are likely to follow that will in turn lead to drop in hardware costs and more evolved AI solutions around the same. All said and done this use case seems to be some distance away from becoming mainstream. Despite a huge potential, getting reliable face recognition solution based on hardware + AI software to mass learner population will take time. Meanwhile, software-only AI based solutions that use device cameras can be a good substitute as interim solutions.
Accessibility Considerations
In very simple terms, face recognition relies on taking an image of the face using a camera. One essential element in the process is the need for the user to face the camera. Wherever possible to implement with reliability, it is recommended to enable face recognition with accessibility considerations taken care of as guided by Sec 508 and WCAG 2.1 AA standards.
Face Recognition and Privacy Laws (COPPA & FERPA for Education)
A ton of questions are being asked today on Face ID reliability and legality. Reliability is bound to get better as AI training models improve and hardware gets better. Legality will need to be addressed in all situations. Apple seems to have covered themselves for now on both these grounds (though time will tell). Face map is classified as biometric record and hence personally identifiable information (or PII). If the student is a minor, sharing of the record is subject to COPPA / FERPA regulations (and potentially other state legislations). There are some genuine privacy concerns. Solutions relying on face recognition will need to ensure face identification data is secure. Apple confirms they do not transmit biometrics to the cloud and the information never leaves the phone. Other solutions will have to build similar restrictions to safeguard privacy and security.
With advances in technology, reliable and affordable solutions are on the anvil. Education is set to reach out to learners with new possibilities. Face recognition will further the cause and is set to become another channel for enabling personalized education.

Ashu Bhatnagar

Head of Delivery at Magic Software, is engaged in building exciting new content and technology solutions for education.