GPT Image 2.5 照片转撕纸拼贴
提示词把上传照片上半部保留写实、下半部转手工纸撕边，并让元素延伸成纸雕浮雕，配中文诗句与英文小译。
GPT Image 2.5 Prompt: Keep the uploaded photograph photorealistic and intact in the upper half of the composition. Below the midpoint, transition into a sheet of warm cream handmade paper with an irregular, naturally torn edge. The entire lower half should remain mostly blank, with a subtle fibrous texture. Allow one recognizable element from the original photograph to extend across the torn edge, transforming seamlessly into a three-dimensional paper-craft relief with delicate layered details and tiny projecting paper ledges. Include 2–3 miniature human figures interacting naturally with this element, creating the feeling of a meticulously crafted miniature diorama. In the open space on the lower paper, write a short, elegant two-line Chinese poem in delicate handwritten calligraphy. Directly beneath it, add a small, understated English translation in soft grey handwritten lettering. Keep both texts beautifully spaced, legible, and harmoniously integrated into the composition. Style & Atmosphere: Surreal photo-to-paper collage, refined paper engineering, handcrafted textures, subtle three-dimensional depth, miniature diorama artistry, soft natural daylight, delicate shadows, macro-level detail, understated elegance, and a calm, contemplative, poetic mood. Composition: Vertical 3:4 aspect ratio. Preserve the original photograph in the upper half, create a clearly defined yet organic torn-paper transition around the midpoint, and leave generous negative space across the lower half. Maintain a sophisticated editorial aesthetic with realistic materials and seamless visual continuity between photography and paper sculpture.