Design, Development, and Quality Evaluation of A Tensorflow.Js-Based Multiple Face Recognition System For Campus Gate Enrollment Verification

Multiple Face Recognition, Tensorflow.Js, Enrollment Status Verification, Biometric Authentication, ISO/IEC 25010, Liveness Detection, Gate-Based Verification, Technology Acceptance Model, Educational Institutions.

Authors

  • Kyle Andrei Domingo School of Science and Technology, Centro Escolar University – Manila, Philippines
  • John Guiller Hernando School of Science and Technology, Centro Escolar University – Manila, Philippines
  • Eliza B. Ayo School of Science and Technology, Centro Escolar University – Manila, Philippines
June 26, 2026

Downloads

Manual student enrollment verification at institutional entry gates is susceptible to human error, identity misrepresentation, and administrative inefficiencies — particularly during high-volume entry periods. This study describes the design, development, and evaluation of the Multiple Face Recognition System (MFRS), a browser-based, gate-deployed biometric platform developed using TensorFlow.js and face-api.js for real-time student identity verification and enrollment status validation in higher education institutions. Grounded in the Technology Acceptance Model and the Systems Development Life Cycle, the MFRS was built using a descriptive-developmental research design and implemented with React.js, Node.js, Express, and PostgreSQL. Core system features include simultaneous detection and recognition of up to three faces per scanner request, Euclidean-distance facial matching against a centralized template database, active head-turn liveness validation, role-based access control for administrators and security personnel, and automated timestamped enrollment verification logging. The system was evaluated through technical quality assurance testing, expert assessment using ISO/IEC 25010 software quality criteria (n = 5 technical evaluators), and a student respondent acceptance survey (n = 50). Results indicated an overall technical evaluation mean of 3.99 (Excellent) and a student acceptance mean of 4.32 (Excellent), with Usability rated Outstanding (M = 4.50). Backend API performance averaged 6.37 milliseconds per response under controlled conditions. Overall recognition accuracy reached 77.05% under controlled conditions, while the active liveness validation success rate was 53.85%, indicating the need for further anti-spoofing development before full institutional deployment. The MFRS constitutes a functional prototype with a clearly defined improvement pathway for gate-based biometric enrollment verification in Philippine higher education.