White Paper Overview
Facial recognition is transforming how we manage access and track attendance — not just in high-security environments, but also in everyday applications. In this white paper, TQ-Embedded presents the results of an in-depth study on implementing biometric face recognition using embedded hardware. Leveraging deep neural networks trained on synthetic datasets, and deployed on the TQMa93xxLA module powered by the Ethos U65 NPU, the study demonstrates how efficient, real-time AI inference can be achieved even on compact, low-power systems.
In addition to performance benchmarks, the white paper explores spoofing countermeasures through presentation attack detection, offering insights into how facial recognition can be both reliable and secure. Whether used as a second layer of authentication alongside traditional access cards or as a standalone method for non-critical systems like elevators or time-tracking terminals, this paper lays out practical use cases and implementation strategies. Download now to discover how TQ-Embedded is making facial recognition smarter, faster, and more accessible.