Start by using an image-to-image AI model and remember this four layer formula: Subect, Context, Style and Technical - S.C.S.T.
As they say, a picture is worth a thousand words. Your image will convey your prompt and you can simply add items like text or instructions such as upscale, reduce aging, etc..
"A realistic hyper-detailed vintage motorcycle, centered, exact silhouette match to input reference, standalone vector style artwork, solid fill design, sharp crisp edges, high contrast ink layout, no bleeding, pure absolute transparency, no drop shadows, no black canvas background container, un-cropped standalone apparel graphic element, 4K resolution 300dpi PNG format"
Match background coin image with realistic 3D metallic texture, detailed coin engravings, sharp metallic reflections and typical coin edge, perfectly centered composition, isolated on a flat dark gray background. Add bas-relief engraving texture, die-struck shallow depth-relief, straight down nadir view, perfect circular boundary geometry, 90-degree orthogonal camera angle, high contrast highlights.
