Fast, very lightweight algorithm for camera motion detection?

Question

I\'m working on an augmented reality app for iPhone that involves a very processor-intensive object recognition algorithm (pushing the CPU at 100% it can get through maybe 5 fra

Answer 1

there's an open source (free for private use) project that utilizes FAST corner detection here: http://www.hatzlaha.co.il/150842/FAST-Corner-V2

It could be very useful for object detection, and has undergone serious optimizations to reach beautiful non-jittery results.

-- EDIT --

Now there's a Lucas-Kanade Optical Flow project as well - http://www.success-ware.com/150842/Lucas-Kanade-Detection-for-the-iPhone You can download the source code, and there's a link to the AppStore as well, so you can play around with it and see if it answers your needs.

HTH,

Oded.

Answer 2

Why not use the combo of the accelerometer/gyro motion sensing, and a very low res image tracker? Each method seems to be confused by completely different user motions.