26. 힉스필드 _ AI 인플루언서.png

https://youtu.be/dw5FKUeXET8


[선착순 20명] AI 수익화 강의 200만원 넘어서 포기했다면? AI 영상 수익화 특강 11기 마감 전 신청하기👇

[선착순 20명] AI 영상 수익화 원데이 특강 11기 사전 신청서



힉스필드 Genjutsu : AI 인플루언서 제작 프롬프트

30대 인물 생성

	@[이미지 1](image_1) 를 @[이미지 2](image_2) 과 같은 구조의 캐릭터 시트를 만들어줘

BAD 챌린지 영상 생성

In the `TASK` section, replace the final paragraph with:

Everything else stays exactly as in @[동영상 1](video_1): every dance movement, pose, gesture, head movement, hand position, foot placement, weight shift, turn, body angle, timing, rhythm, character position, framing, camera behavior, and audio cue. Preserve the source performance energy, but adapt the facial expressions naturally to the facial anatomy and relaxed appearance of @[이미지 1](image_1). Natural facial identity takes priority over literal frame-by-frame copying of the source dancer’s facial deformation.

Add the following paragraph to the `@[이미지 1](image_1)` description under `ACTIVE REFERENCES`:

Use the large close-up portrait in @[이미지 1](image_1) as the primary facial-quality reference. Preserve his naturally relaxed eyes, softly positioned eyebrows, smooth forehead, subtle cheek volume, closed and gently curved lips, balanced jawline, and calm, approachable expression. The front and side views define the correct three-dimensional facial structure. Do not treat the reference expression as a rigid mask; allow subtle, anatomically natural expression changes while preserving the same identity.

Replace the entire `CHARACTER CONSISTENCY` section with:

FACIAL IDENTITY AND NATURAL EXPRESSION LOCK

The replacement dancer’s facial identity must remain stable and photorealistic from the first visible frame to the last. Use the close facial portrait in @[이미지 1](image_1) as the primary reference for the eyes, eyebrows, nose, lips, cheeks, jawline, ears, skin texture, hairline, and resting facial proportions.

Retain the recognizable facial structure of @[이미지 1](image_1), but do not freeze the face into a static portrait. The facial muscles respond subtly and naturally to the dance, breathing, balance changes, head turns, and playful mood.

Preserve the source dancer’s overall confidence and positive energy without copying extreme facial distortions literally. Reinterpret each expression so it fits the replacement character’s facial anatomy. Favor restrained, believable expressions: a relaxed neutral face, a subtle closed-mouth smile, a slightly broader natural smile when appropriate, and brief mild surprise or concentration only where supported by the source performance.

Keep the mouth relaxed and proportionate. The lips open only when naturally required by the source action and never stretch excessively. Teeth appear only briefly and naturally when the smile genuinely opens. Do not create an oversized grin, permanently open mouth, unnaturally wide lips, frozen smile, warped teeth, duplicated teeth, dark empty mouth, protruding tongue, or exaggerated jaw movement.

Keep both eyes aligned and alive, with natural focus direction, small coordinated eye movements, and occasional brief blinks. The eyelids move together with the eyes and cheeks. Do not create glassy eyes, crossed eyes, mismatched pupils, wandering gaze, asymmetrical blinking, oversized eyes, frozen eyelids, or sudden changes in eye shape.

Eyebrow movement remains subtle and coordinated with the eyes and forehead. Cheek movement, smile lines, and nasolabial folds appear gently when appropriate and relax afterward. Avoid excessive eyebrow lifting, deep artificial wrinkles, inflated cheeks, collapsed cheeks, rubber-like skin, or a waxy expression.

The jaw, chin, cheekbones, nose, forehead, and ears maintain their correct three-dimensional proportions during front views, three-quarter views, profile turns, head tilts, and fast movement. Do not stretch, compress, rotate, or reshape individual facial features independently.

During fast turns or motion-heavy frames, preserve the face through natural source-matched motion blur instead of attempting to redraw exaggerated facial details. The face must transition smoothly between angles without flicker, morphing, sudden sharpness changes, or temporary replacement by the original dancer’s face.

Skin must retain natural pores, fine texture, gentle tonal variation, and soft facial shading. Avoid plastic skin, excessive smoothing, beauty-filter effects, waxy highlights, painted texture, harsh sharpening, green color contamination from the chalkboard, or visible face-swap boundaries.

The hairstyle and hairline remain exactly consistent with @[이미지 1](image_1). Do not age the character, add facial hair, change his ethnicity, alter his facial proportions, or blend his identity with the original dancer.

In the `5.0–10.0s` section, replace “changes in facial expression” with:

subtle and naturally adapted facial-expression changes that preserve the relaxed facial anatomy and recognizable identity of @[이미지 1](image_1)

Replace the first two paragraphs of the `QUALITY` section with:

Stable character and background rendering throughout the entire approximately 22.3-second video. The replacement dancer stays on-model in front view, three-quarter view, profile view, fast motion, bent poses, and partially cropped positions.

Facial expressions remain subtle, fluid, symmetrical where anatomically appropriate, and consistent with the natural face shown in @[이미지 1](image_1). Preserve stable eyes, pupils, eyelids, eyebrows, nose, lips, teeth, cheeks, jawline, ears, skin texture, and hairline. No rigid portrait mask, frozen smile, exaggerated grin, overextended mouth, warped teeth, glassy eyes, mismatched pupils, facial asymmetry, feature sliding, rubber-like skin, face-swap seams, or frame-to-frame expression flicker.

Add the following items to `POSITIVE LOCKS`:

- Use the close-up portrait in @[이미지 1](image_1) as the primary facial-quality reference.
- Preserve body choreography exactly, but adapt facial expressions naturally to the replacement character.
- Prioritize stable identity and believable facial anatomy over literal copying of exaggerated source expressions.
- Keep expressions relaxed, subtle, fluid, and appropriate to the dance.
- Maintain natural eye focus, coordinated blinks, restrained eyebrow movement, realistic cheek motion, and proportionate mouth movement.
- Use natural motion blur during fast movement instead of distorted facial reconstruction.
- No frozen face, oversized grin, permanently open mouth, warped teeth, glassy eyes, plastic skin, facial feature sliding, or identity flicker.

50대 인물 생성

@[이미지 1](image_1) 인물이 50대 일 때 모습으로 변경해서 동일한 구조의 캐릭터 시트를 만들어줘

한강공원 배경 이미지 생성

@[이미지 1](image_1)  구도만 유지 후 나머지는 한강 공원으로 변경

격투신 영상 생성

TASK
Edit @[동영상 1](video_1). Replace the hooded monk fighter with @[이미지 1](image_1), using the complete facial identity of the young East Asian man, his light skin tone, neatly styled short black hair with a soft center part, lean build, black textured button-front cardigan, white crew-neck T-shirt, loose black trousers, and black leather lace-up shoes exactly as shown in @[이미지 1](image_1). Replace the attacking martial artist in the white gi with @[이미지 2](image_2), using the complete facial identity of the middle-aged East Asian man, his light-to-medium skin tone, short swept-back salt-and-pepper hair, lean build, black textured button-front cardigan, white crew-neck T-shirt, loose black trousers, and black leather lace-up shoes exactly as shown in @[이미지 2](image_2). Nothing of either source fighter’s face, body, clothes, footwear, necklace, sash, gi, belt, armor, or wraps carries over.
Replace the complete source dojo environment with the modern open-air concrete riverside pavilion in @[이미지 3](image_3) throughout the entire video. Remove the source wooden plank floor, red lanterns, rear shrine dais, red railings, source columns, calligraphic signs, decorative artifacts, and all other source set dressing not visible in @[이미지 3](image_3). Stage the duel on the level gray tiled floor, with both fighters supported by the tiled surface during every step, crouch, landing, hand brace, and fall.
Everything else stays exactly as in @[동영상 1](video_1): every body movement, pose sequence, martial-arts action beat, character position, camera move, framing change, continuous-shot rhythm, and timing.
ACTIVE REFERENCES
@[이미지 1](image_1) is the complete appearance reference for the defensive fighter: use the young East Asian man’s exact facial identity, neatly styled short black center-parted hair, light skin tone, lean build, black textured button-front cardigan, white crew-neck T-shirt, loose black trousers, and black leather lace-up shoes.
@[이미지 2](image_2) is the complete appearance reference for the attacking fighter: use the middle-aged East Asian man’s exact facial identity, short swept-back salt-and-pepper hair, light-to-medium skin tone, lean build, black textured button-front cardigan, white crew-neck T-shirt, loose black trousers, and black leather lace-up shoes.
@[이미지 3](image_3) is the mandatory full-location replacement: use its spacious modern open-air concrete riverside pavilion, broad gray tiled floor, exposed concrete columns and overhead beams, horizontal metal louvers, open river view, distant arched bridge, city skyline across the water, riverside railings, trees at the far right, and warm late-afternoon sunlight casting long shadows across the floor.
SOURCE VIDEO
@[동영상 1](video_1) is the source video being edited. Preserve the opening airborne high kick over the defensive fighter’s low duck, the recovery into guard, the advancing punch flurry, forearm parries, spinning strike, brief hand-to-floor recovery, the second high kick, the defensive duck and spinning counter strike, the attacker’s backward collapse, and the winner’s controlled rise to stand over the fallen opponent. Keep the single continuous moving camera shot and its real-time pacing exactly in place.
SHOT-BY-SHOT GUIDE
Shot 1 (0.0-10.1s): Keep the same eye-level medium-to-medium-wide tracking and orbiting view of the uninterrupted fight. The @[이미지 1](image_1) fighter begins in the left foreground in three-quarter profile, crouched low on the gray tiled floor with both forearms raised in guard and his gaze fixed toward the attacker. The @[이미지 2](image_2) fighter begins in the right midground facing left, launches the same flying high kick over @[이미지 1](image_1), rotates through the landing, and returns to fighting range on the tiled floor.
From 2.8-5.2s, keep @[이미지 1](image_1) shifting backward through the left and center foreground while he deflects the incoming punches and spinning strike with his forearms; keep @[이미지 2](image_2) advancing from the right toward center with aggressive forward punches, then briefly dropping a hand to the tiled floor to catch his balance after the failed attack.
From 5.2-8.2s, keep @[이미지 2](image_2) throwing the committed high roundhouse from the right side toward center; @[이미지 1](image_1) ducks beneath it, pivots on the tiled floor, and delivers the same decisive spinning counter strike, sending @[이미지 2](image_2) recoiling backward.
From 8.2-10.1s, keep @[이미지 2](image_2) falling flat onto his back in the center-to-right midground with arms spread and remaining down on the tiled floor. Keep @[이미지 1](image_1) recovering into a low crouch near center-left, then straightening to stand over him with a calm downward gaze.
Show the target pavilion’s gray tiled floor beneath both fighters, its exposed concrete columns and overhead beam structure around them, and the open river, bridge, and distant city skyline in the background at the correct scale and depth. Preserve the warm sunlight and long natural shadows visible in @[이미지 3](image_3). Do not show the removed plank flooring, lanterns, shrine steps, red railings, source signs, tatami mats, shoji screens, timber structure, or any other element absent from @[이미지 3](image_3).
SETTING
Use the complete modern open-air concrete riverside pavilion from @[이미지 3](image_3) as the only environment. The fighters occupy the open central tiled floor, scaled naturally beneath the high concrete ceiling and exposed structural beams. Keep firm foot contact, knee contact, hand contact, and full-body landing contact with the gray tiled surface.
Preserve natural depth through the foreground floor tiles, central open fighting space, repeated concrete columns, overhead beams, riverside railing, broad river, distant arched bridge, city skyline, and trees at the far right. Maintain the warm late-afternoon light entering through the open sides and the long column shadows extending across the floor. Do not add weapons, furniture, decorations, walls, or architectural elements that are not visible in @[이미지 3](image_3).
EXPRESSION NOTES
Keep @[이미지 1](image_1) focused and composed through the evasions and counter, then calm and victorious while looking down at the fallen opponent. Keep @[이미지 2](image_2) intense and aggressive during the attacks, shocked at the counter impact, then stunned and defeated as he falls and remains down on the tiled floor.

용산공원 배경 이미지 생성

@[이미지 1](image_1)  구도만 유지 후 나머지는 용산 공원 거리로 변경

제품 홍보 영상 생성

Replace the existing `@[이미지 2](image_2)` paragraph under `ACTIVE REFERENCES` with:

@[이미지 2](image_2) is the character reference: a young East Asian man with light skin, a clean-shaven face, a lean build, and the exact hairstyle shown in the reference. His hair is short and naturally black, with a softly voluminous top, a gentle center-to-slightly-off-center part, and smooth curtain-style front sections that separate naturally above the forehead. The sides are neatly tapered around the ears without being shaved, and the back remains short and clean. Preserve the exact hair length, silhouette, parting position, volume, strand direction, hairline, side profile, and natural matte texture visible in @[이미지 2](image_2).

He wears a black textured button-front cardigan over a plain white crew-neck T-shirt, loose black trousers, and black leather lace-up shoes. Face, hairstyle, build, outfit, footwear, and identity 100% match the reference throughout — he is the only person in sharp focus. The visual appearance of @[이미지 2](image_2) takes priority over any general hairstyle interpretation.

Do not add a cap, earrings, necklace, bag, layered long sleeves, cargo pants, or any clothing or accessories absent from the reference. Do not restyle, lengthen, shorten, recolor, slick back, spike, curl, wet, flatten, or exaggerate his hair.

Insert the following new section immediately after `ACTIVE REFERENCES`:

HAIR CONSISTENCY LOCK

The character’s hairstyle must match @[이미지 2](image_2) exactly from the first frame to the last. Maintain the same short black hair, softly raised crown volume, natural center-to-slightly-off-center part, separated curtain-style fringe, neat side length, clean back, hairline, and overall silhouette.

As the camera descends and completes the full 360-degree circle, reconstruct the hairstyle accurately from every angle using the front, side, and close facial views in @[이미지 2](image_2). The front view defines the parting, fringe separation, hairline, and facial framing. The side view defines the crown height, side length, ear exposure, taper, and back silhouette.

Allow only subtle natural strand movement caused by the outdoor air. The hairstyle must not change shape during the camera orbit or the character’s toast. No identity drift, hairstyle drift, changing part direction, expanding hair volume, shrinking hair volume, disappearing fringe, uneven side length, bald patches, duplicated strands, or flickering edges.

Do not generate a bowl cut, buzz cut, exposed undercut, shaved sides, long hair, heavy bangs covering the forehead, sharp side part, slicked-back hair, spiky hair, strongly curled hair, wet-look hair, dyed hair, gray hair, or any hairstyle not visible in @[이미지 2](image_2).

Replace the `QUALITY` section with:

QUALITY

Stable rendering at constant real speed: he stays on-model from every angle of the circle. His facial identity and exact short black center-parted hairstyle from @[이미지 2](image_2) remain unchanged during the overhead view, descent, frontal view, profile views, rear three-quarter views, and final frame. Preserve the same hair length, crown volume, fringe separation, hairline, side taper, back shape, and matte black texture without flicker or restyling.

The can stays exact — silver-white body, green top band, circuit graphics, AI emblem, lime graphic, proportions, and label sharp and unwarped whenever it faces the camera. The crowd stays smooth in the blur with no flicker, popping, or duplicated people. No ghosting, duplicated limbs, hairstyle deformation, facial drift, or façade warping during the descent and orbit.

Replace the character sentence in `POSITIVE LOCKS` with:

He matches @[이미지 2](image_2) exactly from every angle, including the complete facial identity and the precise short black hairstyle with its soft center-to-slightly-off-center part, naturally separated curtain fringe, controlled crown volume, neatly tapered sides, clean back, and unchanged silhouette. He is the only person in sharp focus; the crowd stays at the edges, softly blurred, and never crosses between the camera and the character.