Styling Split and Merge

In this video, we'll look at two sibling nodes: Styling Split takes one picture apart into four pieces, and Styling Merge puts the pieces back together.

When you add Styling Split from the Add node menu, LUVI asks: Wire the styling rig?

Press Wire it and the rig is set up for you; press Cancel and you keep just the node.

In the rig, one source picture feeds four Seedream generators, each dropping its result into its own Media node.

The split's four outputs carry four separate prompts: identity, wardrobe, location and pose.

Its menu settings, Population, Fat density and Muscle density, shape the identity lane only; both densities run from zero to twenty.

We drop a photo into the source Media node and run it: four generations at one hundred and forty-four credits each.

The identity lane gives the person on black, the wardrobe lane the clothes laid out one by one on white, the location lane removes the people, and the pose lane moves the stance onto a grey mannequin.

Styling Merge does the reverse. Press Wire it, and four role pictures go into an ordered Stack, four role instructions into the merge, and both meet in one Seedream generator.

The role instructions arrive as Prompt nodes you can edit, and a role you leave unconnected stays out of the instruction.

We drop identity, wardrobe and location from the split, and the pose from the previous video's cook, into the Media nodes in order: one generation, one hundred and fifty-nine credits.

The merge's menu shows the instruction it wrote: the pictures are numbered Image 1 to 4, and the Stack carries them in that same order.

The result: the same person, in the same outfit, on the same street, in the cook's stance. In short: the split breaks a picture into four roles, and the merge joins them in numbered order.