Linear interpolation for image upscaling

I am currently working on a linear interpolator which will be used to increase the resolution of images in real time. The input images can be of arbitrary size and the output is fixed. The pixels arrive (8 bit per pixel) at the input one at a time.

I planned to use the basic algorithm c*u[n] + (1-c)*u[n+1], c=[0,1], but can't figure out a good way to calculate c (or the samling positions). Do someone have experience of such a problem or just a good idea?

Thanks
Martin

martin.wahlstedt