Abstract:

Capturing and maintaining students' attention in a classroom can be a difficult task for any teacher, regardless of their years of experience. Factors such as the subject being taught, the nature of the students, and the working conditions can contribute to this challenge. However, in a distance learning setting, this problem becomes even more significant, as the teacher has little or no control over what happens on the other side of the students' screens.
This work proposes the use of an avatar for virtual classes. An avatar is a virtual representation of a person. Despite their undeniable success in the entertainment industry, very few documented cases exist where such technologies are applied for educational purposes. Consequently, there are no commonly compatible applications of virtual avatars with the diverse virtual classroom platforms available (such as Zoom, Google Classroom, Microsoft Teams, etc.).
To make this idea possible, we present an application of virtual avatars that teachers can use in their online classes. With the help of this application, it is expected that teachers will be able to improve students' attention during online classes, as students interact with a virtual representation of the teacher that reacts to their gestures and movements using a motion tracking algorithm called Face Tracking.
VTeacher allows teachers to use an avatar within their preferred education platform to help maintain consistent student attention and provide a unique online learning experience. The tool utilizes an optical recognition algorithm to track the teacher's movements and integrates NDI (Network Device Interface) technologies for effective video transmission. Furthermore, VTeacher allows for the incorporation of custom virtual avatars to enhance the educational experience.
VTeacher was designed to be as straightforward as possible so that teachers only need to open the application and share its window on their preferred educational platform. This way, the avatar, clearly following the teacher's movements, would be visible to the students, while the necessary teaching materials are displayed in the background.
VTeacher utilizes a Face Tracking algorithm that detects and locates human faces in a video and extracts information about their movements to be replicated by an avatar. The video is retrieved in real-time from the camera of a desktop computer or a mobile device such as a cellphone. Additionally, it includes a Unity-based engine for the execution of customizable avatars.
The engine generates a virtual space where the avatar is presented, reacting to the teacher's face tracking to replicate their movements. An initial test of this component can be seen at https://www.youtube.com/watch?v=orhetblofM8.
VTeacher utilizes NDI (Network Device Interface) technology to ensure optimal integration of the teacher's desktop with the avatar. NDI is a high-performance standard developed by NewTek, which enables the use of low-latency real-time video in existing IP video networks.
The application is functional and managed to captivate all individuals to whom the project was presented. In its current state, the application successfully validated our hypothesis that both teachers and students believe it could genuinely help increase attention in class. However, it should be noted that further in-depth studies and tests involving larger numbers of students and teachers are required to corroborate this assumption.

Keywords:

Avatar, face tracking, motion capture, e-learning, computer assisted education.