A comprehensive character sheet prompt designed for GPT-Image 2 to generate consistent outfits and layout views for a high-end restaurant floor manager, including front, three-quarter, and back views.
Komut
# Simple Character Sheet Generation Instructions for Sharing
Reference only the single attached full-body person image directly to generate one 1:1 square image. Do not request other people, costumes, restaurants, or layout images.
## Highest Priority: Consistent Person
Use the attached image as the direct editing source for the person, not just for atmosphere. Maintain the face, hair, proportions, shoulder width, torso length, waist position, body type, limb length and thickness, center of gravity, and original posture across all panels. Do not have the person take new poses, correct them to stand upright or symmetrical, and fit the costume to the body. Panel 01 should replace only the costume while keeping the original image's posture; 02, 03, and the face views should keep the person fixed and move only the camera. Naturally complement only the invisible back side, and do not change the person's identity, beautify them, enlarge the head, shorten legs, or standardize the body type. If conditions conflict, prioritize person consistency over layout.
## Costume
floor manager at a high-rise luxury restaurant. A pearl ivory collarless jacket (high-quality matte crepe) with ultra-fine burgundy piping on the front opening, pocket edges, and cuffs; a burgundy inner top; a small burgundy/pearl ivory/soft silver scarf; a modest metal nameplate with no text; a burgundy high-waisted tight pencil skirt ending slightly above the knee; neutral stockings; and black closed-toe low block heels. Must be completely uniform across all panels; do not use a gray inner top.
## Layout
Realistic and high-quality reference material. Use a strict 6-column grid on an off-white square canvas with equal outer margins, card borders, and spacing. Do not create common headers, titles, logos, or decorative lines; start directly with the top panel labels.
- Upper tier: approx. 60%. Three equal divisions (2 columns each).
- Lower tier: approx. 37%. 50%/50% left/right (3 columns each).
- Label strips are bright off-white with a thin warm gray border and neat black sans-serif text.
## Upper Tier
Background is the same bright pale gray. The person should be approximately 88-92% of the image height, including the full body and shoes. Align the scale, crown, waist, knees, and soles across the three figures.
1. `01 PRIMARY REFERENCE`: Same person, posture, and composition as the attached image, changing only the costume.
2. `02 THREE-QUARTER VIEW`: Camera view of the same person and posture from an approx. 45-degree angle.
3. `03 BACK VIEW`: Camera view of the same person and posture from the back.
Do not include the restaurant, furniture, or props in the upper tier background.
## Lower Left
`04 RESTAURANT REFERENCE`
An empty, high-rise luxury revolving restaurant at night. Depict a circular dining area, circular floor inlay lines showing the rotating structure, continuous floor-to-ceiling curved windows, city night view, dark espresso wood floor, subtle brass accents, burgundy chairs, pearl ivory linens, high-quality tableware and wine glasses, and warm indirect and ring lighting. Do not make it too dark; do not include giant central pillars, reception counters, bars, emergencies, or damage.
## Lower Right
`05 FACE REFERENCES`
Divide the area into three equal parts. Arrange the face to upper torso of the same person with the same expression and costume in a single horizontal row across three camera views: base, approx. 45 degrees, and side. Align face size and eyeline, and use the same pale gray background as the upper tier.
Include only the five labels listed above as text in the image. Do not include names, descriptions, additional numbers, logos, or watermarks. Do not add full-body figures to the lower tier, and do not place people inside the restaurant. Generate only one image that satisfies all of the above.
Bu prompt nasıl kullanılır
1
Yukarıdaki tam prompt'u kopyalayın.
2
GPT Image 2'i destekleyen bir platform açın (örneğin YouMind) ve prompt'u yapıştırın.
3
Konuyu, stili veya ayrıntıları fikrinize göre değiştirin ve üretin.
Bu, YouMind'ın prompt kütüphanesinden ücretsiz bir AI prompt'udur. Kopyalayıp uyarlayabileceğiniz binlerce görsel prompt'u daha keşfedin.
Yapay zekanın on binlerce komutu taramasına izin ver. Modele, tarih aralığına ve anahtar kelimeye göre filtrele; görüntülenme, kaydetme, paylaşım gibi etkileşim metriklerine göre sırala.
Her fotoğrafı ayrıntılı bir AI görsel prompt'una dönüştürün. Ücretsiz görselden prompt'a dönüştürücü kompozisyonu, stili ve ışığı analiz eder; istediğiniz görünümü saniyeler içinde yeniden yaratırsınız.
