Upload an image to use as the visual base. The model analyzes this image’s structure, shapes, and key details, then applies the selected template on top of it. For best results, use clear, well-lit images with the main subject visible.