A conceptual prompt strategy for GPT-Image 1.5 focusing on achieving artistic output by controlling information density. Step 1 involves increasing information (dynamic angles, complex poses) and Step 2 involves reducing information (omitting background details, focusing on precise facial features).
Step 1: Increase information Eliminate horizontal and vertical lines and make various parts of the picture diagonal. Increase variations in joint twists, composition, poses, camera tilt, etc. Step 2: Reduce information Thoroughly erase information that doesn't need to be seen. Draw faces precisely and omit others.
Generate opens this prompt in the editor, ready to run.
Everything after the generation, on the same site.
Same model, same dataset — GPT Image 1.5 prompts people have shared.