Monitoring the Student's Entry and Exit Time in the Classroom

Monitoring the student's entry and exit time into a classroom is a crucial task. This system will be a solution for performing that crucial task. In this system, for face detection and face recognition, two pre-trained models were imported and utilized to create face embeddings for a collection of photographs kept in a file directory. The face embeddings and corresponding labels are then used to train an SVM classifier using the scikit-learn library's SVC class. To detect and recognize the students, the pre-trained face detection and recognition models were loaded using OpenCV's dnn module. Then the saved face embeddings and labels were also loaded using pickle, which was previously generated by a face recognition model. Face detection is performed using a Caffe model trained on the SSD framework, while face recognition is done using OpenFace. The system identifies the student's faces using embeddings that were generated by pre-processing the images and were saved in the face_embeddings.pickle file. It then matches these embeddings to the corresponding labels that were saved in the labels.pickle file. Then the input is captured from the webcam and the optical flow analysis is applied to detect motion in the video stream. After that, it initializes the input image with a blurred, grayscale image and sets the starting points for the motion tracking. It then loops over the captured frames, blurs and converts them to grayscale, and uses calcOpticalFlowPyrLK to calculate the new points based on the old ones. If the new points are within a certain range, it draws a line between them on the input image and displays it. If the points move outside a certain threshold, it triggers an event to log the entry or exit of a student from a classroom and appends these details into a text file accordingly. Finally, this text file is sent through email to a specified list of recipients (faculties) on a daily basis.