AI-BASED FACE RECOGNITION ATTENDANCE SYSTEM

The system is implemented in Python with three main scripts: ·                           collect_data.py: Captures images via webcam. Input: prompts user for student name. Process: Detects face frames and saves multiple images into a folder named after the student. Output: a directory dataset/Name/ filled with face images. ·                           encode_faces.py: Processes the collected dataset. Input: reads all images under dataset/. Process: Detects faces and computes encodings for each image. Output: a file encodings.pickle containing lists of 128-D encoding vectors and corresponding names. ·                           recognize.py: The live attendance loop. Input: webcam stream and encodings.pickle; user-entered subject name. Process: For each video frame, detect faces, encode them, compare with known encodings. If a match is found, check the CSV for duplicates, then append a new attendance record (Name, Subject, Date, Time) to attendance.csv. Also display identified names on the video feed. Output: updated attendance.csv file and on-screen results. A helper script fix_dlib.py may be included to resolve common library issues (on some systems). Files written/used: - encodings.pickle: Binary file of known face embeddings. - attendance.csv: Log of attendance records (Name, Subject, Date, Time). - .gitignore: to exclude large data and pickles. All scripts use libraries: OpenCV (opencv-python), face_recognition (with dlib), and NumPy. No code is shown here, but the logic flow is as described.

[1]             F. Schroff, D. Kalenichenko, J. Philbin, “FaceNet: A Unified Embedding for Face Recognition and Clustering,” CVPR, 2015[1][13]. [2]             J. Deng et al., “ArcFace: Additive Angular Margin Loss for Deep Face Recognition,” [3]             CVPR, 2019[2]. [4]             Y. Taigman et al., “DeepFace: Closing the Gap to Human-Level Performance in Face Verification,” CVPR, 2014[3]. [5]             “face_recognition: Recognize Faces in Images,” Python library documentation (Adam Geitgey)[4][9]. [6]             OpenCV Documentation: Face Detection using Haar Cascades[5]. [7]             V. Suresh et al., “Facial Recognition Attendance System Using Python and OpenCV,” Journal of Software Eng. & Simulation, vol.5, no.2, 2019. [8]             R. Jadhav et al., “AI-Powered Facial Recognition Attendance System Using Deep Learning and CV,” IJRSI, vol.XII, no.IX, 2025. [9]             S. Smitha, P.S. Hegde, A. Rashid, Face Recognition based Attendance Management System, Yenepoya Inst. Tech., 2016.