The problem of automating the students’ attendance in the classroom is solved by using computer vision. A convolutional neural network is used to recognize a person’s face. The recognition process is implemented in real time. Localization of faces in frames from a video camera is performed by the Viola-Jones method. The convolutional neural network of the VGGFace model forms the features of a person’s face. Identification of the person occurs by the facial features similarity. The software is implemented by using the Keras and OpenCV libraries. The control system performs the following functions: captures the faces of students on a video camera when entering the classroom, compares faces with a database of students, notes the presence at the lesson (or being late) in case of successful identification, saves the data in attendance register. For the convenience of video monitoring, the color of the student’s line changes depending on his condition: not present, present, late, absent. The system provides for manual editing of the electronic register and the choice of the subject name. The student’s photo can be uploaded into the database from a file or directly from the camera.
CITATION STYLE
Fedyaev, O. I., & Tkachev, N. M. (2021). Operational Visual Control of the Presence of Students in Training Sessions. In Studies in Computational Intelligence (Vol. 925 SCI, pp. 262–268). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-60577-3_31
Mendeley helps you to discover research relevant for your work.