Biometric and smart identification reader (B.A.S.I.R.): combating proxy attendance via multiprocessing dual-factor authentication

The increasing reliance on single-factor attendance systems, such as Quick Response (QR) codes and Radio Frequency Identification (RFID), has enabled widespread proxy attendance in higher education institutions. This issue is supported by a survey of 250 respondents conducted at the Centre for Foundation Studies (CFS), International Islamic University Malaysia (IIUM) and Universiti Tun Hussein Onn Malaysia (UTHM), where 94% of respondents reported witnessing QR code sharing and 84% of lecturers suspected proxy attendance without verifiable proof. This work aims to develop an attendance system that eliminates proxy manipulation while maintaining real-time performance in high- volume academic environments. The proposed solution, Biometric And Smart Identification Reader (B.A.S.I.R.), integrates RFID-based identification with facial verification and liveness detection. The system utilizes DeepFace (Facenet512) for 1:1 facial matching and MediaPipe for Eye Aspect Ratio (EAR)-based blink detection to ensure user authenticity. To address latency issues commonly associated with biometric systems, multiprocessing is implemented to distribute computational tasks across multiple Central Processing Unit (CPU) cores, enabling continuous operation at 30 frames per second (fps). Experimental evaluation through a fraud simulation test demonstrates that B.A.S.I.R. reduces proxy success rates to 8.2%, compared to 89.5% for QR codes and 36.4% for standalone RFID systems. Authentication latency is reduced from over 5.0 seconds to 1.4 seconds per user, supporting high-throughput operation without queueing delays. In conclusion, B.A.S.I.R. achieves a 91.8% integrity rate while maintaining real- time performance and demonstrates strong commercial potential for secure attendance management.