Writing effective prompts for text-to-image tools like CSM’s Text to Image, Midjourney, or ChatGPT can be challenging. Here are key tips to help you generate high-quality images optimized for Image to 3D:
Game Assets
When creating stylized props or characters for game engines, images often look flat. Use keywords like:
3D asset of a… to achieve greater depth.
Game asset of a… for clear, game-ready visuals.
Stylized or Lowpoly Assets
To generate cartoon-like or simplified objects, include phrases such as:
Lowpoly asset of a…
Lowpoly game asset of a…
A low-poly game asset illustration of a…
Getting Clear Angles
For optimal results, specify clear and readable angles to ensure the best mesh quality. Useful keywords include:
Front view
¾ view or Quarter view
Isometric View
Back view
Object Parts on a Sheet
ChatGPT excels at generating images with separated parts, useful for assembly in 3D:
…with all parts laid on a sheet for Image to 3D.
…create all major parts on a sheet, for Image to 3D-based assembly, I want it in 3D, no spurious parts, only major components.
Specify angles for parts to improve depth perception:
The cylinder should be at a slight angle for better 3D mesh generation.
For a comprehensive guide, see our detailed tutorial here.
Images with Solid Backgrounds
Use clean backgrounds to isolate your object clearly. Helpful phrases include:
…on a white background
…on a black background
…on a solid color background
