Click to skip the navigation bar
  FAQs

How to optimize facial recognition performance of Tapo Security Device

Facial recognition is a powerful tool that adds a new layer of intelligence and convenience to your home security. By optimizing your camera's setup and understanding its capabilities, you can significantly enhance its ability to correctly identify familiar faces and alert you to strangers.

Optimize camera placement

The physical placement of your camera is the foundation for effective facial recognition.

It is crucial to angle the camera so it captures approaching individuals head-on, with a tilt angle of no more than 30 degrees. Avoid side-view angles, as they provide fewer facial details.

Provide adequate lighting, but avoid overexposure

Lighting conditions play an equally critical role. The system works best with clear, well-lit images. You should avoid pointing the camera directly at a strong light source, like the morning or evening sun, as this backlighting can make faces appear as dark silhouettes.

For nighttime recognition, please avoid placing the camera near surfaces such as walls, eaves, or dense foliage. These surfaces can reflect the camera's built-in infrared (IR) light, causing glare and overexposure. This forces the background to darken, making it difficult to recognize faces in the distance. An unobstructed view is key to clear night vision performance.

Understand recognition limitations

Facial recognition is most effective at a distance of 2-3 meters (7-10 feet), although devices with higher resolution can often detect faces from further away. Please be aware that recognition can be challenging if a face is partially obscured by a mask or hat. Success is also less likely if a person moves too quickly, does not face the camera, or works during poor weather conditions, such as heavy rain or strong lens flare.

Keep software updated

Finally, to ensure you benefit from our latest advancements, please ensure that your camera/doorbell/hub's firmware is always updated to the latest version. We regularly release software improvements that refine our algorithms and enhance the performance of features like facial recognition.