Researchers have developed AdaEye, a system for periocular recognition that offers a more efficient alternative to full-face verification, particularly when faces are occluded or image resolution is low. The study systematically scaled input and backbone parameters for periocular verification in both visible-spectrum (VIS) and cross-spectral (VIS/NIR) settings. Results show that direct periocular processing significantly reduces computational cost and increases throughput, though its effectiveness varies by evaluation protocol. The research also found that moderate backbone width reductions can maintain performance while decreasing computation, and embedding compression helps reduce template storage requirements. AI
IMPACT This research offers a more efficient approach to biometric verification, potentially reducing computational costs and storage requirements for AI systems used in security and identification.
RANK_REASON The cluster contains an academic paper detailing a new system and its evaluation. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Hugging Face Daily Papers →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →