Prompt template for creating more details (use this in a VLM with the image):
Analyze the provided image and write a clean visual description to replace [Details here] in the prompt below.
Base prompt:
Convert this image to realistic version
Your task:
Describe only the visible details of the subject and the scene in a way that helps convert the image into a realistic version.
Rules:
- Focus on physical appearance, clothing, accessories, pose, expression, body structure, materials, and environment.
- Describe the scene/background only if it is visibly relevant.
- Do not mention art style, medium, anime style, cartoon style, illustration style, rendering style, or image quality.
- Do not explain the character’s story, identity, franchise, or personality unless it is directly visible.
- Keep the description clear, direct, and natural.
- If the subject is a human or humanoid person, describe them realistically.
- If the subject is a creature, monster, animal, or non-human being, DO NOT humanize it.
- Preserve the original species and anatomy of non-human subjects.
- Do not turn monsters into humans or “human-like versions” unless the image already shows them that way.
- If the creature has unusual anatomy, body proportions, extra limbs, animal features, or a non-human face, keep those details explicit.
- If skin tone or colors are affected by lighting, describe the apparent color carefully without overcorrecting.
- Output only one final paragraph to be used as [Details here].
Output format:
[Details here]
