From faces to HDR scenes objective and perceptual metrics for AI smart glasses image quality

Artificial Intelligence (AI) smart glasses with integrated cameras are becoming increasingly prevalent, yet their image quality remains underexplored. This study presents a comprehensive evaluation protocol tailored for such human-facing wearable devices, using both standardized and proprietary metrics (including AI based metrics). Key metrics include Local Contrast Gain (LCG), face exposure, texture preservation, visual noise in Just-Noticeable-Difference (JND) units, Video Exposure Convergence (VEC), and video stabilization across static and walking conditions. Perceptual evaluations were also conducted to assess HDR performance in high (D65, 1000 lux) and low (2700K, 5 lux) illumination environments. The Meta Ray-Ban AI smart glasses were benchmarked against an iPhone 16 Pro Max smartphone using the Ultra wide camera and several other AI-enabled smart glasses including the Xiaomi AI Glasses, Rayneo V3, and Meta Oakley HSTN to evaluate their strengths and limitations in real-world AR/VR use cases.