Testing Partial Image Correction and Pose Modification in GPT-Image-1.5

The user attempts to modify a generated image by keeping the person's appearance the same while changing their position and context (moving them to the passenger seat of a left-hand drive car, conversing with the driver). The test showed mixed results regarding position and context adherence.

Prompt

241 charactersOpen Astorie studio ↗
■I want the person in the reference image to be in the passenger seat of a car without changing their appearance.
・The car should be {argument name="car side" default="left-hand drive"}
・Appearing to be smiling and conversing with the driver
From sourceView source ↗

More to explore

Structured Selfie Portrait Prompt — output

Find your next direction

ASTORIE CANVAS

Take the idea further.

Bring your prompts, references and results together in one creative space.

Open your canvas ↗