For normal HDR as I understand it, an intermediate image is created, based on image data from a set of images with varying dynamic range. The intermediate image has a narrower dynamic range than the HDR image, but is generally more faithful to the real thing. The intermediate image is used to compare different tone mappings and colors for creating the final image. At least, that's how I learned it. Is that not what you do? Since I don't like the look achieved by HDR most of the time, I don't use this technique very often.
For image fusion, you just merge overlapping images to quickly blend differently exposed images into an output image. I used to have to do this manually, but there are several computer programs available now that will do it. It has the benefit of maintaining natural colors and doesn't produce any halos. So it's more similar to what we used to do in the darkroom to blend multiple negatives into a single print. This technique can also be used to expand the DOF in an image, by blending focus stack of multiple images of the same subject, taken at different focal lengths or resolution. We use this technique with my work, when analyzing satellite or aerial imagery and other data into a single image or map (monochromatic satellite images frequently have sharper resolution than color, so you combine the two and presto!).
As for using this technique on film images that have been scanned, I don't see why not. The scanned image you create is just an image file same as any other. You'd want to scan your images in RAW format for the best results, though.
