Latest & Greatest

[#1]

Ever wonder how the facial recognition software on digital cameras functions?  If so, then you may want to check out Allan-BBY's explanation of the feature from our Cameras & Camcorders board:

Hi Everyone,

How often have you tried to capture an image of someone with your camera just to have your camera focus on something other than your subject? If you’re anything like me you probably feel like you have lost the opportunity to capture a special moment in your life, and you wish your camera had done a better job of focusing on your subject.

Before autofocus the decision of what to focus on was entirely up to the user, but this often meant the moment in time had passed before the image you saw could be captured. Thankfully autofocus systems were developed, and now almost always work much faster and more accurately that manually focusing on your subject.

There of course are exceptions to the above statement, and sometimes these autofocus systems focus on something other than your subject. This is often due to your subject not being in the center of the frame, or there is something closer to you than your subject.

Camera manufacturers have been working on overcoming this limitation, and now manufacturers have what is called Face Detection technology that in most cases overcomes this limitation. My wife’s Canon G9 has this technology, and I can attest to the fact that it focuses on faces in the frame the vast majority of the time.

When I first used this technology I couldn’t understand how it could possible detect faces in a scene, but after looking into how this technology works I was surprised how logical of an approach these manufacturers took.

Due to contrast differences caused by lighting three dimensional shapes the manufacturers developed advanced algorithms that search for common facial features of humans in the frame, and then shift focus to those points. Without the processing power of modern camera processors this technology would be impossible due to the time it would take for less advanced processors to run these algorithms.

This technology of course has to work within the limits of the camera, camcorder, or webcam using it. Because of these limits there are certain situations where it won’t be able to properly focus on all the faces in a particular frame. For example if have one person three feet from the camera, and your second subject is twenty feet away, this technology won’t be able to focus on both unless the lighting and camera lens make it possible to achieve the depth of field necessary to focus on both subjects.

It isn’t a perfect technology, and there are definitely situations that even this technology can’t overcome, but all-in-all I feel like it’s definitely worth having.

I hope this helps!