100 Images with Only Four Base Images

This post describes a workflow for creating as many images as possible from only four base images.

The setup is simple: we have one scene, one person, one product, for example sunglasses, and that product should be photographed on the person inside an environment.

What we need

To do that, we need:

• the product sheet for the sunglasses
• the person with the outfit
• the face sheet with the sunglasses on, ideally as a precise reference
• the environment, if needed

For studio environments, an environment image is not absolutely necessary. For other scenes, however, an environment image is still recommended.

How the sheets are created

These sheets are created with separate techniques for each sheet:

• sunglasses
• person with outfit
• face with sunglasses
• environment

Once we have these sheets, they are all fed into one node and given to the AI model as reference images.

The AI model then builds the final image by combining all of the sheets into one result.

Why the sheets matter so much

A short but important point: these sheets are extremely important. They matter even more than the final prompt itself.

The reason is simple: the sheets are the context for the AI model.

They tell the model what the person looks like, what the product looks like, and what the environment looks like. Without that context, the result becomes chaotic. The image will look different every time, the product will look different every time, and the result will not feel professional.

These sheets are essential.

Workflow links

The workflows are linked here:

• Workflow 1
• Workflow 2