I don't know where image_add() or Image_subtract() came from. They're not part of MATLAB or anything that I can find on the File Exchange.
That said, these are probably the simplest things to replicate. Assuming that the images are the same size and class, you can use imadd() and imsubtract() from the Image Processing Toolbox. You can also just literally add or subtract the numbers using + or -.
If the class or number of channels may vary, you'll have to manage that accordingly before you add/subtract them. Imblend() from MIMT (on the File Exchange) can do add/subtract blends regardless of a mismatch of class or number of channels. That said, it's probably severe overkill for just that.
