I am working o开发者_高级运维n a project that needs to do face detection on a camera video input (like security camera).
I managed to open Video input via AVFoundation framework. I managed to capture a UIImage from the video input, then feed into opencv library to do face detection.
But the process of capturing requires at least 2 to 3 seconds.
Is there anyone who got some experience to share ?
If you can work with the limitation if iOS 5, you can get this to work pretty well with the new CIFaceDetector
baked into the OS. That should help address performance concerns and simplify the code if you're not worried about implementing your own version of face detection.
Apple has some sample code projects up on the developer forums here. There are some details on the kind of data available using their face detection here and some basic docs for the detector here.
精彩评论