Will Your Camera Recognize Faces?

Sometimes even the most seasoned experts struggle to figure out if the equipment is set up right when it’s first installed or after a year of use.

Installing cameras outdoors comes with its own set of challenges. You might need to mount a camera higher to keep it out of reach, or the perfect spot for identification could be way farther than expected, making it tough to focus the lens manually. Plus, there’s always the possibility of missing how glare from the sunrise, sunset, streetlights or advertisements could mess with your view.

And a year down the line? Well, things change. People’s movement patterns shift, branches from trees might start blocking the view, lenses get dirty, the camera’s sensor starts to fade, or the camera tilts from wind, vibrations, or birds.

After working on dozens of face recognition projects for both outdoor and office cameras, we realized there was a serious need for a tool to help us, our contractors, and clients figure out whether their cameras were properly calibrated and set up.

So, we decided to create that tool. We tested it, and let’s just say—we’ve seen its value firsthand. The truth is, there are 14 factors that determine how well face recognition works. And our report helps you understand and tackle 10 of those factors head-on.

Internal factors:

1. Equipment stability
2. Network bandwidth
3. Camera resolution
4. Camera sensor quality
5. Video analytics server performance (overload >80%)
6. Quality of reference photos in the database

External factors:

1. Vibrations (from wind, traffic, etc.)
2. Weather conditions (snow, rain, fog)
3. Distance between the camera and the object
4. Pedestrian traffic density
5. Speed of movement within the frame
6. Backlighting (from sun, streetlights, advertisements)
7. Camera angle (head tilt and rotation)
8. Lighting conditions (faces lit with less than 200 lux)

Omni Agent detects faces in video, tracks them for a few seconds to select the best shot (based on size, angle, sharpness, lighting, and more), and saves it to the Omni Platform database.

When we generate a report, the faces captured by your cameras are analyzed and examined against 19 different parameters using our very own QAA algorithms, based on NIST standards. In the end, you get a detailed, easy-to-understand assessment of your camera’s performance—showing you not only what’s going wrong but also how to fix it.

And we even pinpoint problems that would be hard to spot with a simple visual check. Impressive, right?

1. When assessing the possibility of using existing equipment

Some clients say, "We have 1,000 cameras. Can you tell us which ones can be used?" And you're left thinking, "Do I really need to check each camera? How do I even verify this?"
Offer your client a project for evaluation and complete it automatically with the 3DiVi service!
2. Identify efficiency loss over long-term use
On long-term projects, detecting a loss in camera performance is crucial. Use the 3DiVi service to pinpoint these issues and confidently offer maintenance and support services to keep everything running smoothly.