Using convolution to upscale an image

I understand how write code to apply a convolution matrix to a pixel. :slight_smile: Not that hard. Now each upscaling algorithm can be implemented using a convolution kernel. I am just not sure how to apply this to code. Lets say I have a 1000x1000 px image. And I upscale it to 1500x1500 px. Now each pixel maps to 1.5 pixels. How exactly do you apply your convolution in code ?

Could someone kindly show a sample code or pseudocode algorithm for this?

I will start by asking in stackoverflow after doing a detailed google research on this topic. A convolution matrix to upscale an image is too generic and even if you show a reference or document about your algorithm, it is very likely the code has already been written out there. Share it if you find any and that would be a better start to discuss this further.


1 Like