RGB-D camera-enabled digital twin dataset for education

Official URL: https://dx.doi.org/10.1109/SIU66497.2025.11112313

The concept of a Digital Twin originated in 1997, but its foundations date back to the simulations created by NASA for space missions in the 1960s. Although it did not gain significant attention compared to other fields of computer science for many years, Digital Twin technology has experienced a resurgence in interest, particularly following the COVID-19 pandemic, which highlighted the importance of online education. By identifying students' focal points, we can enhance the quality and efficiency of education. It is essential to simulate such scenarios within digitally created environments, like labs and classrooms, and analyze these scenarios based on where and for how long students focus their attention. To achieve this, the Intel RealSense D435i, a modern RGB-D camera, was utilized to collect real-world data for the digital environment. This camera's depth measurement capabilities enable the accurate digital representation of labs, classrooms, and meeting rooms, facilitating the testing of real-world scenarios. The point cloud data obtained from the camera was processed using MeshLab and imported into the Unity environment for animation creation. These animations will be capable of detecting where students are looking in real time, resulting in the compilation of a "gaze line"dataset that illustrates where and for how long students focus their attention, ultimately enabling the development of realistic animations. The data and appendix will be available at: https://7unahan.github.io