Pocket City Model

model

gpt-image-2

Use each uploaded photo as the only visual source.Create one independent vertical 3:4 poster per photo and output each separately.Never combine uploads.Divide the canvas into two strictly equal horizontal zones:the original photograph fills the top 50%,and a white presentation panel fills the bottom 50%.Keep the midpoint boundary fixed.If needed,crop the source naturally to fit the upper half without stretching or changing the subject. Top photo:Preserve the uploaded image as a realistic photograph.Keep the main subject,objects,lighting,color,depth,and camera viewpoint unchanged.Apply subtle grading only.Do not repaint or add graphics. Bottom miniature:Reconstruct the most recognizable subject or scene from the photo as a highly detailed 3D miniature on the white lower panel.Preserve the core identity,main buildings or landforms,major objects,and key colors,but do not rigidly copy the crop or viewpoint.Allow the miniature to use a more open display angle,including a mild three-quarter view,so it reads as a complete model.Naturally extend cut-off edges by completing partially visible buildings,roads,shorelines,stairs,trees,or nearby structures with believable continuation based on the source.Do this minimally,so the place still feels the same.Avoid abrupt truncation of important forms.Add one neat regular display base beneath the miniature,like a clean architectural plinth.The base should be thin,orderly,and geometric,with clear edges.It should support the miniature as one unified object without becoming a box or case. Scale and placement:Center the miniature horizontally in the lower half.Make its width about 42-46% of the full poster width,with 44% as the target.Place its center around 74% of the full canvas height.Leave ample white space around it. Caption:Add one English-only location caption near the bottom center,at 92-94% of canvas height.Use only one line of English text in uppercase or small caps,with no Chinese or other non-English text.Use the exact English place name or transliteration supplied by the user.If no reliable place name is available,use a short scene-based English title instead. Typography:Use a refined classic serif with light regular weight and wide tracking.Keep the caption modest,clean,and centered,in soft charcoal.Do not add a second line or native-language title. Overall style:Premium travel-object poster and collectible miniature presentation.The upper half remains photographic;the lower half stays spacious white with one miniature on a tidy base and restrained captioning.Keep the transformation clear and elegant. Negative:No stylized upper photo,no stretched source,no oversized miniature,no chopped-off main building,no floating object,no messy irregular ground slab,no extra props,no decorative frame,no busy lower panel,no Chinese text,no large headline,no bold title,no fake location,no random text,no logo,no watermark,no UI,no cartoon,no anime,no painterly rendering.

parameters

Reference Images