
ChatGPT
DesignChatGPT
Photo to funny crayon scribble comparison poster
Creator @Hamburgerai
Prompt
Please turn every photo I upload into an individual 3:4 vertical transformation comparison poster, no multi-image stitching, output each photo separately. The screen adopts a clear top-and-bottom dual-link structure: the upper half retains the complete original photo, and the lower half transforms the same subject into a rough, funny, slightly grimace-like crayon scribble. The upper part maintains the identity, quantity, posture, movement, clothing, animal fur color, architectural structure, environmental relationship, observation angle, and real lighting of the subject in the photo, with only slight photographic color grading. Do not add or delete people or animals, do not change the location, and do not turn the photographic part into an illustration. The lower part retains the most recognizable outline, posture, and relationship of the subject in the original photo, but do not perform realistic copying. Refine and appropriately enlarge the faces of people or animals, deliberately exaggerating eyes, eyebrows, nose, teeth, tongue, corners of the mouth, and facial proportions to form a skewed, asymmetrical, and slightly clumsy funny grimace. Expressions can be more absurd and dramatic than the original photo, but must still be recognizable as the original subject; do not turn into horror images, and do not apply fixed cartoon characters. Use thick black crayons, oil pastels, and oil pencils to repeatedly search for outlines. Lines should be shaky, broken, overlapping, uneven in thickness, and retain traces of erasing, back-strokes, coloring outside the boundaries, and untrimmed marks. Extract colors from the original photo and compress them into a limited color palette, representing bodies, clothes, or hair with rough friction, local white space, and uneven overlapping colors. Do not use smooth outlines, clean color blocks, or exquisite digital coloring. The background only retains a few clues sufficient to identify the location or action relationship, such as building outlines, seats, railings, roads, trees, or travel landmarks, and summarizes them with more scribbly, low-density strokes. The subject should be more concentrated and impactful than the original photo, with a part of warm paper white reserved around it; do not redraw all the photographic details fully. Add a set of short English handwritten text in the blank space of the lower half: - Refine a 1–3 word main title based on the image; - Add the field "SCRIBBLE STUDY 01"; - Add a short observation describing the real action or expression in the image. The text uses skewed, dry, slightly broken pencil or crayon handwriting, like a painter's casual notes on the edge of the paper, clearly legible but not overly regular. No author names, brands, URLs, logos, watermarks, fake quotes, garbled code, or redundant text should appear. Overall presentation is a deliberate non-pretty manual graffiti feel: rough, funny, absurd, with a bit of a grimace quality, while retaining the real character relationships, animal characteristics, and travel memories in the original photos. Avoid: exquisite cartoons, cute style, Japanese anime style, smooth vectors, realistic sketches, horror and gore, uniform template faces, complete background copying, excessive decoration, changing the number of subjects, changing character postures, fictional landmarks, and turning the photographic part into an illustration.
Copy the text, add your own details, and try it in your image model.