4 Tips to Help You Optimize Your VMS with Integrated Face Recognition System
At 3DiVi we understand the intricacies involved in setting up and configuring a Video Management System (VMS) with Face Recognition capabilities.
Here are four essential tips to help you optimize your VMS setup, ensuring accurate, reliable, and efficient face recognition performance:
Prioritize camera focus over resolution: Instead of opting for high-resolution cameras, choose specialized long-focus cameras with lower output resolution that can capture large faces in the frame. This approach reduces network bandwidth requirements, storage needs for video data, and server capacity needed for video analytics.
Edge processing is key: Process video data at the edge (directly at intersections where cameras are installed) using specialized edge devices, rather than transmitting the full video stream to a central data center. This minimizes the risk of identification losses due to datatransmission failures, reduces costs associated with building and maintaining communication lines and networking equipment, and decreases the storage requirements in the datacenter.
Maintain high-quality database photos: Ensure that the original photos in the database have a quality score above 80% (you can use this tool for evaluation). Lower quality photos increase the likelihood of false identifications or misses.
Set a high identification threshold: For operational systems, the recommended identification threshold should be above 87.6% to reduce the number of false positives in databases with more than 500,000 faces. Security services do not have the resources to handle false identifications, so the principle of "if it's not recognized by this camera, it will be recognized by another" helps to minimize false alarms and maintain the system's credibility.