CVOCV: A CoreVideo-based OpenCV experimentation environment.
CVOCV was built in order to create an extremely simple and lightweight OpenCV experimentation environment for OS X 10.5.x. Using CoreVideo makes it possible to send video frames to OpenCV with a minimal amount of computational overhead.
In order to build this project, you must be running OS X 10.5 Leopard.
The XCode project can be downloaded here (Updated 1/24/09):
http://buza.mitplw.com/buzamoto/CVOCV_b003.zip
Also find it on github:
Looking for an iOS version of this? Look no further! (Updated 3/15/11)
Documentation
Documentation can be found in the downloaded zip.
Motivation
With only a single machine on which to run all of my applications, when I need to accomplish specific tasks, particular attention must be paid to performance. When I wrote this example, I was looking for a simple way to do OpenCV-based blob detection, but in a way that minimize CPU overhead. Existing approaches typically exist embedded within other infrastructures that make specific assumptions about the intent of the programmer. I needed a simple way to get blobs, and after finding them, to send the results to another application that is able to more appropriately handle its task.