Smart Attendance System Using a Hybrid Face Recognition Model

Though often overlooked, tracking who shows up matters deeply in schools. Yet many places still rely on slow hand-written logs, demanding constant oversight. These old ways open doors to tricks like someone signing in for a friend. Lately, though, attention has shifted toward solutions rooted in body-based checks - fingerprint scans, facial patterns - to handle sign-ins automatically. What boosts performance isn’t only pace - fewer openings for errors play a key role too. A fresh method for tracking attendance emerges by blending two distinct face identification strategies - MobileNetV2 and LBPH. Rather than depend on one pathway, the framework links neural network outputs with pixel-level patterns. When light swings between dim and bright, most setups struggle - this one adapts faster. Though shadows stretch or vanish, performance holds steady. Where brightness surges suddenly, adjustments follow without delay. Even under flickering change, stability remains high. As contrast jumps, sensitivity tunes itself ahead of drop-offs. Once harsh highlights appear, balance resets within moments. One part draws compact facial summaries using lightweight networks. Texture details rise in clarity when neighboring pixels undergo comparison across regions. Even if someone tilts their head slightly, matches still form reliably. Performance climbs not through complexity, but thoughtful pairing of complementary logic. Testing inside regular classrooms shows the tool works with far better precision and reliability. Noticeable is how light the build feels, which helps when moving it around.