Which AI models work best with this prompt?

Models that support detailed structured JSON prompts like Sora 2, GPT-based image generators, or similar platforms handle this layout prompt most reliably.

Can I change the subject or setting?

Yes, the prompt includes editable variables for the subject description, setting, and model name. Swap in any portrait or environment while keeping the grid structure intact.

What should the output look like?

Expect a dark vertical poster in roughly 4:5 aspect ratio with a 7×7 grid of 49 tiles—48 labeled FACS action units plus one empty tile—each showing a thumbnail with an anonymized face and a caption bar.

GPT Image 2Prompt AI

FACS Action Unit Reference Grid — Sora 2 Prompt

Generate a sleek contact-sheet poster mapping all 48 FACS facial action units onto a consistent portrait model for reference or creative use.

1 / 2

PROMPT · EN

type

FACS 面部动作编码系统参考网格海报

title

{argument name="headline text" default="FACS 参考网格 — 48 个动作单元 (v3 Step 2/3 修复)"}

subtitle

OLD 算法女性，咖啡馆窗边 | 模型名称：{argument name="model name" default="banana 2"} | 眼睛 v3.1，闪光灯预览

overall_style

深色技术联系表布局，小型白色无衬线字体，细网格间距，摄影缩略图，温暖的咖啡馆窗边光线，哑光黑与炭灰色 UI 背景

subject

description

{argument name="subject" default="留着长黑发、手持米色陶瓷杯靠近嘴边的年轻女性"}

setting

{argument name="setting" default="温暖的咖啡馆窗边内景，带有垂直窗框和金色日光"}

face_treatment

每个缩略图展示相同的肖像构图，面部中心区域覆盖有大面积柔边米色矩形匿名遮罩，同时保留头发、手部、杯子、颈部和背景

layout

canvas

竖版海报，约 4:5 纵横比，黑色背景

header

左上角双行标题，紧凑且具有技术感

grid

count_total_tiles

count_filled_action_units

count_empty_tiles

columns

rows

thumbnail_style

每个方格上方为摄影图像，下方为深色标题栏，包含数字索引、AU 代码和动作标签

empty_tile

右下角方格为空白深灰色，带有微小文字显示 missing 以及标签行 49. ---

action_units

item 1

index

code

AU01

label

内眉上扬

item 2

index

code

AU02

label

外眉上扬

item 3

index

code

AU04

label

降眉

item 4

index

code

AU05

label

上眼睑提升

item 5

index

code

AU06

label

面颊提升

item 6

index

code

AU07

label

眼睑收紧

item 7

index

code

AU09

label

鼻皱

item 8

index

code

AU10

label

上唇提升

item 9

index

code

AU11

label

鼻唇沟加深

item 10

index

code

AU12

label

唇角拉升

item 11

index

code

AU13

label

唇部尖锐拉升

item 12

index

code

AU14

label

酒窝形成

item 13

index

code

AU15

label

唇角下压

item 14

index

code

AU16

label

下唇下压

item 15

index

code

AU17

label

下巴提升

item 16

index

code

AU18

label

唇部噘起

item 17

index

code

AU20

label

唇部拉伸

item 18

index

code

AU22

label

唇部漏斗状收缩

item 19

index

code

AU23

label

唇部收紧

item 20

index

code

AU24

label

唇部挤压

item 21

index

code

AU25

label

双唇分开

item 22

index

code

AU26

label

下颌下垂

item 23

index

code

AU27

label

嘴部拉伸

item 24

index

code

AU28

label

吸唇

item 25

index

code

AU41

label

眼睑下垂

item 26

index

code

AU42

label

眯眼

item 27

index

code

AU43

label

闭眼

item 28

index

code

AU44

label

斜视

item 29

index

code

AU45

label

眨眼

item 30

index

code

AU46

label

单眼眨眼

item 31

index

code

AU51

label

头部左转

item 32

index

code

AU52

label

头部右转

item 33

index

code

AU53

label

抬头

item 34

index

code

AU54

label

低头

item 35

index

code

AU55

label

头部左倾

item 36

index

code

AU56

label

头部右倾

item 37

index

code

AU57

label

头部前伸

item 38

index

code

AU58

label

头部后仰

item 39

index

code

AU61

label

眼球左转

item 40

index

code

AU62

label

眼球右转

item 41

index

code

AU63

label

眼球上转

item 42

index

code

AU64

label

眼球下转

item 43

index

code

AU65

label

斜视（外斜）

item 44

index

code

AU32

label

咬唇

item 45

index

code

AU33

label

鼓腮

item 46

index

code

AU36

label

舌尖突出

item 47

index

code

AU35

label

吸腮

item 48

index

code

AU19

label

吐舌

generation_instructions

在每个填充方格中创建一致的参考网格，使用相同的主体、摄像机角度、杯子、光线和背景。仅根据每个标签改变面部动作或头/眼方向，即使面部大部分被匿名遮罩隐藏。保持标签清晰易读，并使用准确的 AU 代码和编号。

Informazioni su questo prompt

Generate a sleek contact-sheet poster mapping all 48 FACS facial action units onto a consistent portrait model for reference or creative use. Use it as a Concept Art starting point for GPT Image 2: keep the visual structure and style constraints intact, then swap in your own subject, brand, or scene.

This prompt is organized as structured JSON with fields such as type, title, subtitle, overall style, subject, layout, and action units. Keeping that hierarchy helps the model separate the subject, layout, interface elements, and detail constraints.

Start with small changes to subject, scene, camera, and color, then generate a few variations to compare composition and detail without turning the page copy into keyword stuffing.

Come usare questo prompt

Copy the full prompt JSON into your preferred AI image generator that supports structured prompts.
Choose or confirm the model name variable (default is "banana 2") and set the headline text if desired.
Customize the subject description and setting variables to match your creative vision.
Generate the image and review the 7×7 grid layout with labeled AU codes.
Adjust variables and regenerate to refine lighting, anonymization mask opacity, or grid density.

FACS Action Unit Reference Grid — Sora 2 Prompt

Informazioni su questo prompt

Come usare questo prompt

FAQ

Altro da GPT Image 2

Minimalist Mother's Day Wallpaper — Sora 2 Prompt with Calligraphy & Carnations

Minimalist Skincare Brand Manual — Sora 2 Prompt

Dynamic Nike Ad Poster — Sora 2 Prompt with Sydney Sweeney

Crea più velocemente con i prompt Sora2Hub.