Need Explanation of Bicubic Interpolation, in imresize inbuilt function of MatLab?

I have a Image and I want to resize it by factor of 0.5 in C code . So In matlab I can do by using inbuilt function imresize(Image,0.5 'bicubic') .But I have to convert this imresize inbuilt function in C code. So I need expalnation on how imresize function is working in matlab??

Can anyone please Explain me How Can I processd??

Answer 1

[DGM]

If you're out to replicate imresize(), see this answer:

https://www.mathworks.com/matlabcentral/answers/888177-how-does-imresize-antialiasing-work

and this comment:

https://www.mathworks.com/matlabcentral/answers/1633410-source-code-of-imresize-s-bicubic-interpolation#comment_1946990

Also:

https://stackoverflow.com/questions/26823140/imresize-trying-to-understand-the-bicubic-interpolation

I also recall another conversation that I can't find yet. I'll add it if I do. It might have been deleted.

Ultimately, you'll find that the core issue will be the black box that is imresizemex.mexa64. I don't know what's in that.

If you just want any example of 2D interpolation, there are probably more examples around.

[Annu Sheoran]

@DGM sir Matlab is using bicubic interpolation using Convolution algorthim.

I understand in imresize how weights and indics are calculated but not able to understand after that how we can apply these wights on pixcels .

For example

my matrix is

a = [10,20;30,40]; scaling factor 2,method bicubic

weights genrated by matlab

[-0.0234375000000000 0.226562500000000 0.867187500000000 -0.0703125000000000

-0.0703125000000000 0.867187500000000 0.226562500000000 -0.0234375000000000

-0.0234375000000000 0.226562500000000 0.867187500000000 -0.0703125000000000

-0.0703125000000000 0.867187500000000 0.226562500000000 -0.0234375000000000]

final output using imresize

[7.1875 10.1563 16.0938 19.0625

13.1250 16.0938 22.0313 25.0000

25.0000 27.9688 33.9063 36.8750

30.9375 33.9063 39.8438 42.8125]

how they are gettinh this output im not understanding

Can you please help??

Answer 2

[Image Analyst]

Did you try

>> edit imresize

[Annu Sheoran]

I have to convert Imresize in C code

[Image Analyst]

Then see the algorithm here:

https://en.wikipedia.org/wiki/Bicubic_interpolation

and write that in C code. We won't do that for you since we're only about MATLAB code here.

Answer 3

[Steven Lord]

If you have access to MATLAB Coder, the imresize function supports the C / C++ Code Generation extended capability (with some limitations.) See the Extended Capabilities section on its documentation page for more information about what's supported.

[Annu Sheoran]

@Steven Lord I have to follow misra c standard