Revision history [back]

I have tested the GaussianBlur function of version 3.0(Android binding) in both INTEL and ARM CPUs but I got different results where those are supposed to be identical. For any kernel size for one time blurring ,the difference produced in results is less than or equal to one i.e at some pixels I observed zeros and at remaining pixels I observed ones.This error is getting higher and higher when I tried to run the function in a loop.I used 8 bit RGB images for testing. The remaining image blur functions are working perfectly(giving identical results).But it seems to be an issue with Gaussian Blur.

I have tested the GaussianBlur function of version 3.0(Android binding) in both INTEL and ARM CPUs but I got different results where those are supposed to be identical. For any kernel size for one time blurring ,the difference produced in results is less than or equal to one i.e at some pixels I observed zeros and at remaining pixels I observed ones.This error is getting higher and higher when I tried to run the function in a loop.I used 8 bit RGB images for testing. The remaining image blur functions are working perfectly(giving identical results).But it seems to be an issue with Gaussian Blur.