Photo-to-Isometric: Reality Slice Generator

{
"prompt": "Design an architecturally accurate isometric diorama based on the uploaded reference image, strictly adhering to the visual information present. Begin by meticulously analyzing the image to identify the predominate architectural style, the overall age and condition of the structures, the dominant building materials, the layout of streets and alleys, and the specific objects, vehicles, and urban density.

Your goal is to recreate this scene as a single, floating city block against a pure white background, while preserving the original atmosphere, proportions, and spatial logic. If mid-rise buildings are present in the image, faithfully replicate their facades, including matching textures, balconies, windows, storefronts, and street elements exactly as seen. Remove any elements not present in the source image.

Apply a strict 45-degree isometric angle to the scene, ensuring that the perspective is consistent and accurate. Utilize a tilt-shift miniature effect to create a sense of scale and depth, while maintaining natural lighting conditions that match the original scene as closely as possible. Implement global illumination to enhance the sense of realism, and employ physically based rendering (PBR) materials to capture the true appearance of the surfaces.

Pay obsessive attention to micro detail, ensuring every element is rendered with the highest level of architectural visualization quality possible. Use clean studio lighting to eliminate shadows and highlights that would not occur in the original scene, creating a consistent and uniform illumination across the entire diorama. Do not include any sky or horizon elements, focusing purely on the ground-level scene.

Please provide your design as a single, high-resolution image with a 1:1 aspect ratio, rendered in a strictly photorealistic style with the highest level of detail achievable."

,
"negative_prompt": "Any invented objects, buildings, or elements not present in the original image, any fantasy or stylized elements, any illustration or cartoon styles, any low-poly or flat-shaded geometry, any fisheye or distorted effects, any surreal or inconsistent details, any objects or props added at random, any inconsistent scaling or proportions",

"aspect_ratio": "1:1",
"style": "Strictly photorealistic",
"quality": "Highest possible"
}

This revised version of the prompt provides more specific instructions, clarifies the desired output format, and defines a clear set of constraints to guide the AI's generation process. The instructions are more detailed and the desired output is explicitly specified as a single image with specific dimensions and style requirements. The negative prompt has been expanded to include more specific exclusions, further guiding the AI's generation process.