Six-Look Beauty Styling Comparison — ChaosRealmsAI · MIT

原文

GPT Image 2 · en

Scene:
A square commercial beauty-or-fashion image designed like a polished campaign proof for Beauty Look Era Switchboard. Keep the camera discipline clear and believable: portrait-oriented within the square, premium studio light, controlled tonal contrast, and enough surrounding negative space to feel like an ad layout without relying on readable typography. The image should look production-ready, with realistic skin, hair, fabric, and makeup rendering that supports identity locking instead of distracting from it.

Subject:
Build the frame around one clearly recognizable person. Before After Board should feel like a specific transformation step that a disciplined comparison layout making transformation logic instantly legible. The face structure, age, body proportions, and general camera position must stay consistent with the same-person premise of Transform one face across distinct beauty directions such as clean-girl, Y2K, runway editorial, bridal, and retro glamour while preserving recognizability.. Styling, wardrobe, grooming, makeup, hair, or pose can evolve, but the viewer must believe it is unquestionably the same subject across the whole theme. Use a premium editorial-commercial tone rather than a casual portrait session.

Important details:
The visual test is identity consistency under change. Preserve stable bone structure, eye spacing, nose shape, lip volume, skin tone family, and believable texture response. Let the transformation be visible through styling choices instead of facial drift. This theme studies editing-workflow, multi-image-reference, photorealism, commercial-mockup through a production-ready visual system. Make the composition clean enough for client review, with professional grooming fidelity and enough detail for salon, beauty, or fashion decision-making. Aim for clear hierarchy, polished craft, and immediately usable campaign logic.

Use case:
Beauty Look Era Switchboard single study 1/5.

Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no face drift, no inconsistent age shift, no extra limbs, no duplicate person, no asymmetric eye placement, no beauty-editorial chaos.

使い方

1. パラメータを含む元のプロンプト全体をコピーします。
2. 被写体、文字、構図を自分の要件に置き換えます。
3. 対応する外部ツールで使用し、利用条件と結果を確認します。

開始前の確認

• 参考画像の要否を確認してください。互換モデルの表示は、そのモデルでの生成実績を意味しません。
• 名前、数値、画像内の文字を確認します。モデルの版で結果は変わります。