Expand Image Composition

Scenario

You need to adapt existing images to new ratios for social covers, hero sections, or banner placements without changing the original subject.

Steps

1. Start with a sharp, centered source image.
2. Request target ratio and "change-only" constraints.
3. Generate missing background areas only.
4. Validate seam blending and lighting continuity.

Parameter Guidance

• Preferred ratios: 16:9, 3:1, 4:3 depending on placement.
• Constraints:
  • preserve face, pose, clothing, and camera angle
  • generate only new side areas
• Avoid list: no extra people, no symbols, no random objects.

Failure Troubleshooting

Symptom: Subject changes during expansion

Fix:

• Add stronger identity lock statements.
• Reduce stylistic modifiers.

Symptom: Visible seam between original and expanded regions

Fix:

• Request smooth tonal transition and consistent light direction.
• Re-run with simpler background instructions.

Compliance Boundary

• Do not use expansion to fabricate deceptive evidence.
• Follow policy limits for sensitive identity edits.

FAQ

Can we add text inside generated image area?

For production workflows, keep text overlays in design tools after generation for better control.

How many expansion variants are recommended?

Three variants are usually enough to select one production candidate.