E-commerce-Grade Product Photography: Qwen-Image-2.0 Image Generation Prompt Paradigm

Using a matte-finish skincare serum bottle as the subject, this demonstrates how structured prompts can bring text-to-image up to product photography standards usable for e-commerce detail pages. The approach splits the prompt into four layers: subject and materials (bottle material, pump metal finish, label typography), set and environment (acrylic surface, same-tone backdrop, prop proportions), lighting (key light position, softbox size, rim light and shadow hardness), and imaging quality (focal length, depth of field, micro-contrast, clean commercial retouching feel). Qwen-Image-2.0 supports long instructions of up to 1000 tokens, output up to 2K, and has explicit optimization for Chinese text rendering, so the Chinese product name and volume information on the bottle label can be written directly into the prompt to be rendered verbatim, without needing to add text in post. The example output is a 2048×2048 square image, suitable for the main image slot; the same prompt structure switched to 2688*1536 becomes a detail page banner. Applicable scenarios: e-commerce main images and detail pages, social media ad assets, and batch exploration of first drafts for product visual proposals.