I'm staring to develop my first OpenCV iOS application and I wondering if it is better to use AVFoundation directly instead of using CvVideoCamera (that is a wrapper for AVFoundation).
I need the maximum speed.
What do you think?
Not that I have many experience on mobile platforms, but the general way to improve speed during processing is to use as less wrapper functions as possible. However, CvVideoCamera does more then just wrap AVFoundation, it also makes sure that correct variables are created, assigned and maybe other functions are called.
So if you decided to go into the code and call AVFoundation directly, make sure that you do everything that is done inside CvVideoCamera manually yourself in your code.
Thanks for your answer. Where can I find CvVideoCamera source code? Thanks again.
I guess you should download the openCV source code from the repository, then look at the dev forum where exactly this source code is located.
Also you can use GSOC 2012 examples as a starting point. They are using AVFoundation directly instead of using CvVideoCamera as a wrapper.
http://www.code.opencv.org/projects/ios/repository/show/trunk
Asked: 2013-02-27 04:49:41 -0600
Seen: 939 times
Last updated: Apr 15 '13
