Smart Attendance By Using Face Recognition

Maintaining attendance by hand in contemporary classrooms is a laborious activity that is prone to errors and mistakes. To make this process easier to comply with, we suggest a Face Recognition Based Attendance Management System (FRAMS) which makes use of modern facial recognition technology to eliminate the need for manual attendance. This system utilizes computer vision technology where face detection and recognition take place using real time python open cv software while mingling through a comfortable user interface designed by Tkinter. Haar Cascade classifiers are incorporated for face detection while the facial recognition aspect is conducted through an efficient algorithm referred to as Local Binary Patterns Histograms (LBPH) that is resistant to changes in lighting, facial emotions and facial poses.FRAMS works by taking students pictures using a webcam and a dataset is generated from these pictures. This dataset is then used to prepare the face recognition model. During attendance sessions, live faces are compared against the stored database within the system to recognize and verify users. When the system is capable to identify a person, it automatically registers the person’s state using the online spreadsheet which further provides attendance records in a precise manner. The aided system is efficient as it removes the need for attendants roll call, thereby minimizing errors and curtailing proxy attendance.More so, the system was subjected to extensive testing when in use in various situations and its recognition accuracy rate was high in controlled conditions. There were, however, some problems like the differences in lighting and the chances of wrongful identification in the middle of busy pictures. Even so, in light of these challenges, FRAMS is an improvement over the attendee sheets that yesteryear utilized because it’s faster and more dependable. For works in progression, it may be possible to enhance the recognition algorithms to cope with increasing the size of datasets and the difficulty of the surroundings.