A few examples of different computer vision tools running on ios via openCV. There is an issue with performance and/or cpu throttling after about 5 seconds of image processing that needs to be worked out.
tracks only a few keypoints, using the goodFeaturesToTrack function, faster than dense optical flow. Uses lucas kanade algorithm.
This is doing template matching to look for the "santitas" logo (a brand of tortilla chips)
Calculates hough transform on the image, which is a per pixel accumulator over function space (in this case circles, but also has version in opencv for lines)
algorithm from paper: http://infoscience.epfl.ch/record/175537/files/2069.pdf matches to the santitas logo - invariant to many perspective transformations, but more sensitive to scaling
uses farnebeck's algorithm to calculate dense optical flow fields