To take in more students and conduct higher quality education, classrooms must be upgraded for teachers to reach expected outcomes. Currently, computer classes require teachers to spend large amounts of time and energy to manage classroom discipline and grasp each student’s learning status, to ensure that the quality and outcome are both up to standard. Oftentimes, it becomes hard to track if students are deviating from classwork to surf on the internet, play games, or with their phones. Student’s performance and learning productivity can be improved if teachers are able to more accurately and precisely track student concentration level, amount of practice performed, status of progress, and their general emotional response to class materials. Through utilizing an efficient and multitasking AI BOX, the system can carry out real-time edge computing for deep learning-based behavior recognition of students. The automated management can solve many problems associated with teacher’s high investment to create a more intelligent learning environment.
To realize the smart classroom vision of automated classroom management system, Professor Lee, a university lecturer adopts ASRock Industrial’s iBOX-1185G7E powered by 11th Gen Intel® Core™ i7 processor with Intel® Iris® Xe Graphics to create the AI BOX. The software environment uses Ubuntu 20.04.2 LTS as the operating system to implement Intel® OpenVINO™ 2021.3.394. In combination, the system is able to act as an image inference engine, delivering high performance AI inference for education applications with Deep Learning Boost at 4.15 TFLOPS and 8.29 TOPS. With rich IOs, it collects video data from its camera through USB 3.2 Gen2 to perform efficient AI model computing and AI image inference in AI BOX, and outputs to display through HDMI 2.0a interface, efficiently grasping each student’s real-time learning status.
The AI BOX implementation process adopts the Open Model Zoo sample program and each model performs visual prediction of the video data separately, including Gaze Estimation, Object Detection, Human Pose Estimation, and Emotion Recognition models. Through the Gaze Estimation Demo, the system can make all necessary inferences, and observe whether the students are practicing or temporarily resting by referring to the computer screen operation process. Furthermore, the AI BOX can make inferences based on the Object Detection Demo implementation, this allows teachers to know how many students are currently in their seats, thus counting the presence and absence of the students. From the Human Pose Estimation Demo, teachers can determine through the positions/ postures of students how they are carrying out or perhaps feeling about their tasks. Last but not least, the Interactive Face Detection Demo allows for observation of students’ facial expressions to determine their current condition and whether assistance would be needed.
Overall, much time-consuming administrative works such as counting attendance, compiling results, and analyzing statistics can now be entirely automated. The AI BOX provides educators and students a learning space with better customized and quick service. Under transformative times, education can assimilate with new technologies to provide better and more efficient engagement for our future generation.
iBOX-1185G7E enables the smart classroom
*Image source: the classroom images used in the case study are from Professor Shengan, Lee’s article “【AI In Classrooms】Computer Vision for Classroom Learning Management ” (Jun, 03, 2021) posted on MAKERPRO.cc
The AI BOX carries out automated administrative and evaluative services in the classroom through artificial intelligence. This greatly reduces the time needed for teachers to manage a large number of students, and can effectively monitor learning status of students to improve overall quality of class and teaching outcomes.
The automated system allows teachers to better understand students’ learning status and general emotional state, this allows them to provide more appropriate teaching methods to improve learning outcomes. For example, detection of students’ emotional reactions in class, their commitment levels in front of the computer, the length of focus on exercises, number of times on or off their seats, etc. can all be recorded for further analysis and improvement.
With the increasing popularity of computer vision technology, the high-cost equipment is now replaced. The same function can now be achieved through the AI BOX with its high computing power and high-performance capabilities under much lower costs.