My process is straightforward. I find a suitable discarded snapshot and restore it, while looking for elements from other found images to layer in. In the above case, my starting image was 2 x 3 inches. I scanned it at a high resolution and made initial adjustments, cleaning up dust spots and scratch marks and adjusting contrast.

Next, I layered the bride in and sized her to fit compositionally, then, working her into it seamlessly, taking into account lighting and resolution issues (e.g. the bride came from a much larger, sharper image and had to be slightly blurred, also grain patterns had to be matched).
Finally, I applied an even, overall tone to the image, sharpening it slightly, and making final light value adjustments.
I try to leave signs of imperfection in my work, technical, human mishandling, and natural degradation. This is the reality of most vernacular images, after all.