Re: source code for 2d cross-correlation fx?
thanks, i will take a look...The correlation fx uses the FFT so I suppose if starting from scratch with the FFT and knowing the algorithm for correlation, I can make the modifications as needed. Cross-correlation involves multiplying the FFT of one fx by the complex conjugate of the FFT of the second fx (I believe). Since LabVIEW and IMAQ have all these operations, I should be able to work with this. Sincerely, Don
Re: source code for 2d cross-correlation fx?
At 11:11 AM -0400 4/15/04, Don J Roth wrote: I know this is a long shot but does anyone have the c or c++ code to compute the 2d (image) cross-correlation fx? (I know that IMAQ VISION has the cross-correlation routine but I am looking to potentially modify the 2d correlation routine to compute something called non-symmetry coefficient and would need to start with the source code). I know this is a long shot, but http://developer.apple.com/Documentation/Performance/Conceptual/vImage/index.html has built in optimized vector libraries that do convolution (which is just a correlation) on 2 D images. Having messed with these libraries they are REALLY FAST. These have been tuned for the hardware and use the vector acceleration when they can. But the problem is that you have to be on the correct platform. Other than that, I think that gnu has some signal processing libraries that you might use. -Scott
