Eula from Genshin Impact 'Look' Prompt, 3
March 23, 2026. Mudae has a look prompt for each of its character cards. I put that into Serin to see what would happened. Asymmetrical Legwear, Bare Shoulders, Boots, Colored Pupils, Detached Sleeves, Eye Bags, Fake Claws, Fishnet, Gloves, Habit Veil, Hair Streak, High Heels, Large Breasts, Maroon Hair, Maroon Hair Accent, Multicolor Hair, Pale Skin, Pantyhose, Pelvic Curtain, Purple Eyes, Purple Hair, Red Hair, Short Hair, Sidelocks, Skirt, Snowflake on Design, Thigh High Boots, Thigh Strap, Unique Headwear
Generate a comprehensive character + background refinement image based on Image 1. The goal is to create a fully integrated new scene where the character remains in the image, while the background is rebuilt around them to better match their outfit, vibe, and likely setting. The final result must stay one complete, unified image—not a separate background plate, not a cutout composition, and not a disconnected character pasted onto a new scene.
Core Purpose: Use Image 1 to determine:
who the character is
what they are wearing
how they carry themselves
where they would realistically be
what time of day fits them best
what props and environmental details belong around them
Then generate a single cohesive scene that:
keeps the character present in the image
corrects any body/model inconsistencies if needed
preserves identity
improves the background so the whole image feels natural and complete
---
1. Input Rule
Use Image 1 only as the reference.
The character must remain visible in the final image.
Do not generate a background by itself.
Do not make it look like the character was pasted into a new environment.
The final result must look like one natural illustration built together at the same time.
---
2. Character Preservation Rule
The character is still the subject of the image and must remain intact.
Preserve:
overall identity
body proportions
breast size
facial features
hairstyle, hair length, and color
tattoos and their placement
horns, tails, ears, piercings, scars, cybernetics, or other physical traits if present
outfit design unless the request later says otherwise
If the reference image has visual inconsistencies, awkward anatomy, or rendering issues:
refine/fix the model
but keep the same person
do not accidentally create a new face or new body type
The result should feel like a cleaner, corrected version of the same character.
---
3. Background Generation Goal
Build a new environment around the character that matches:
their outfit
their styling
their body language
their implied mood
the kind of place they would realistically be dressed like that
The scene should answer:
Where would this character be?
What time of day fits best?
What props belong there?
What lighting supports both the outfit and the mood?
---
4. One-Piece Image Rule
The final image must remain all in one piece:
character and background must share the same perspective
character and background must share the same lighting direction
shadows must match the environment
colors must feel unified
the ground/contact area under the character must make sense
the environment must wrap naturally around the character
Do not create:
a blank background plate
a separated character render
a collage look
floating character syndrome
Everything should look intentionally composed as one finished artwork.
---
5. Environment Decision Logic
Infer the best setting from the character’s outfit and vibe.
Examples:
Casual wear: apartment, city sidewalk, café, bookstore, mall, park, balcony
Clubwear / nightlife: neon street, club entrance, rooftop bar, lounge district, alley with signage
Swimwear / beachwear: beach, boardwalk, poolside, seaside resort, beach stairs, changing area
Formalwear: ballroom, luxury hotel, event venue, staircase, terrace, gala hall
Sleepwear / pajamas: bedroom, hotel suite, living room, nighttime balcony, cozy indoor setting
Holiday outfits: Christmas market, decorated home, snowy street, Halloween porch, festive venue
Costume / themed wear: environment should support the theme without overpowering it
The chosen environment should feel like the natural home of the outfit.
---
6. Time of Day Logic
Choose the best time of day based on the outfit and mood.
Examples:
Casual / soft looks: morning, afternoon, golden hour
Swimwear / summer looks: midday sun, warm afternoon, sunset
Nightlife / clubwear: evening, late night, neon-lit night
Formal / elegant looks: sunset, evening, indoor luxury lighting
Sleepwear: early morning, evening, late night
Holiday looks: winter morning, snowy dusk, warm evening glow, pumpkin-lit night
Time of day should strengthen the tone of the outfit.
---
7. Prop and Set Dressing Logic
Add props and background details that fit the outfit and location naturally.
Examples:
Bedroom / home: bed, lamps, blankets, rugs, books, mugs, vanity items
Beach / pool: tote, towel, sunscreen, sandals, umbrella, drinks, chair, boardwalk railings
Nightlife: neon signs, street lamps, cars, club ropes, posters, lounge seating
Formal setting: chandeliers, floral decor, polished floors, wine glasses, curtains, architecture
Holiday setting: trees, lights, ornaments, pumpkins, candles, wrapped gifts, snow accents
Props should make the location believable, not overloaded.
---
8. Mood and Lighting Rule
The environment must match the emotional tone of the character and outfit.
Possible tones:
cozy
cute
confident
flirty
relaxed
elegant
festive
sporty
seductive
mysterious
Lighting examples:
warm sunlight
overcast daylight
golden sunset
cool moonlight
neon glow
candlelight
holiday lights
club lighting
Character lighting and environment lighting must match perfectly.
---
9. Character Placement Rule
The character must feel naturally placed in the rebuilt scene:
standing or posed on believable ground
correct perspective
no floating
no weird cropping against props
no object overlap that breaks anatomy
enough open composition space to let the character remain readable
The background should support the character, not swallow them.
---
10. Style Rule
Maintain a high-quality 2D illustration or concept-art style
Keep the environment detailed enough to feel real
Keep the character as the focal point
Preserve readability and visual clarity
Treat the final output like a polished key visual, not a rough composite
---
11. Execution Logic
1. Analyze Image 1 for the character’s identity, outfit, pose, body language, and overall vibe.
2. Preserve the character in the final image and correct/refine any model inconsistencies while keeping them recognizably the same person.
3. Infer the most fitting location, time of day, lighting mood, and environmental props based on the outfit and vibe.
4. Rebuild the background around the character so the entire image feels natural and unified.
5. Make sure character and environment share consistent perspective, lighting, shadows, and atmosphere.
6. Deliver one cohesive image with the character still present inside the finished scene.
7. Use this language: English
매개변수 표시
- 크기
- 512 x 512
댓글