CVBL-CCTV-Face Dataset for Real-World Face Enhancement and Recognition in Surveillance
Read the original on arXiv Computer Vision →The CVBL-CCTV-Face Dataset provides real‑world surveillance images captured by an 8MP IP Dome CCTV camera, paired with high‑quality reference images taken in controlled settings. The study introduces a modular framework that combines face detection, deep‑learning super‑resolution, and identity recognition using L2‑normalized embeddings to address common surveillance degradations such as low resolution, motion blur, noise, and occlusions. Enhancement models are evaluated under three different setups and the system is deployed for real‑time identity verification in unconstrained environments.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.