ChatGPT has a new feature “Image Generation” which allows you to easily transform exist images with a predefined set of prompts.
You then don’t need to prepare the prompts. Rather, all you need is to select an image and a pre-defined “style” which contains the prompt.
Click “Images” from Left Navigation Bar
Select Style and then a local Image to Upload
Prompt (Automatic)
The following Prompt is auto-inserted.
Using the provided image as reference, create a soft, understated image with an art-school aesthetic. If people are present, preserve their identity while using gentle natural light, muted colors, and a relaxed, slightly awkward pose and place them in a simple, everyday setting such as a studio, classroom, or quiet street. Otherwise, treat the scene or object as the subject or subjects, using gentle light, muted colors, and intimate framing.The mood should feel intimate, youthful, and casually artistic rather than polished or glamorous.
Original Picture
Generated Image Result
Much more realistic then the Ghibli style (which looks pure animation/comics)
AI/Prompt Engineering
Post: ChatGPT: New Image Creation (art school)
–EOF (The Ultimate Computing & Technology Blog) —
428 wordsLast Post: Understanding std::transform_reduce in Modern C++


