A group of researchers at Catholic University (UC) has developed an automated student frequency control system that combines advances in facial recognition technology and the quality of images recorded by smartphone cameras.
The application was tested in rooms of approximately 70 students, for 25 classes, through ten systems of facial recognition algorithms that allow counting the number of faces in a digital record and identify who attended the courses.
"Monitoring student attendance is an important management in many educational institutions, but the task of manually scaling and checking the classic present takes time," said Domingo Mery, a professor in the Department of Computer Science Engineering UC.
The researcher, who led the development of the new application along with Ignacio Mackenney and Esteban Villalobos, students of the Research Graduation program, added that the classroom call of 70 students can reach an average of four minutes per day and 80 minutes . in one semester.
"We believe our automated student-based digital face recognition system can contribute to the management of this type of task, as it saves time for teachers and students, and also avoids false class attendance," Mery said.
In the tests performed, the best system used was FaceNet, a deep learning function based method that achieves 95% accuracy in attendance control, with only one image recorded from the smartphone.
For researchers, the developed application could improve the teaching and learning experience in classrooms. However, they estimate that facial recognition is still far from perfect when faced with low quality images.
With this application, the user can take one or several images of the classroom using his smartphone to register all the students present, previous enrollment stage of each student and their respective image in the system.
The algorithm proposed by the researchers detects the faces in the recorded image and recognizes the students present or absent in the order. The application then stores your class's assistance in your system.
The technology can be used in large and small classrooms. Currently, the application is being developed in conjunction with the company BiometryPass.