
The goal of autofocus is to make something in the picture come out sharp. But if you're taking a photo of people, it's not their hands you want in focus. Recently, camera makers have been adding the ability to detect faces in a scene, track them if they move, and optimize both focus and exposure to make everyone look their best. But not all face-detection systems are equal, as I discovered after testing several compact cameras
on patient friends who posed by indoor light, as well as on passersby rushing through Times Square.
How it Works
All face-detecting cameras compare the scene before them to a built-in library of features derived from images of real people, such as the distances between eyes, patterns of light and shadow, and skin colors. So far, no models can identify a face in profile, and they don't function well in low light, such as in bars or candlelit rooms. But they're not easily fooled: In our tests, none were thrown off by variations in skin tone or by accessories like eyeglasses.
Face-Tracking Face-Off
![]() |
Fujifilm Finepix F40fd |
![]() |
Sony Cyber-shot DSC-W80 |
![]() |
Nikon Coolpix S200 |

| regarding | user | just commented |
|---|---|---|
| Chocolate-Chip Science | overlord_of_bla... | yum.....cookies. |
| Farming in the Sky | popoflojo | I completely agree that |
| Phoenix Digs Deepest Trench Yet | avance | phoenix is a neat robot |
| Make A Walking Beast | Dustin2127 | When can I get one of these? |
| The Future of Mobile Computing | Dustin2127 | This is awesome! Im just |

