Hello, I have some problem in imresize function. I use it in my code for resizing the image using bilinear interpolation but it gives different result as compared to the bilinear interpolation done manually using formula.

For the bilinear interpolation, we use 2-by-2 neighborhood method to find the mid value but the interpolation done by imresize gives different result.

Answer 1

[Matt J]

Floating point differences, I'd guess. Or, you are resampling at different locations from those chosen by imresize.

But then, you've left everything up to the imagination. No examples of the differences you mention. No code to let us reproduce them....

Answer 2

[Image Analyst]

Are you absolutely sure you specified bilinear interpolation in the input argument list, because that's not the default - bicubic is . Let's see your line of code.

If you did, then it could depend on where you and they specify the endpoint locations, like right on the center of an edge pixel or somewhere else, or is maybe just rounding error.