大白鹅闯入证件照四宫格

INTRUDER = "A big white goose"
scene:
"The final image is a 2×2 grid layout, which is the highest priority composition rule"
"The four panels are exactly the same size, separated by thin white borders, the four photos fill the entire frame, the overall image ratio is 3:4, and each panel is also 3:4"
"Each panel is a real ID photo: the person from the uploaded reference image, facing the camera, maintaining a standard ID photo pose"
"The four panels tell a continuous story, the reading order is: top left → top right → bottom left → bottom right"
"In this story, the INTRUDER breaks into the ID photo shooting scene, causing increasingly severe physical chaos to the person"
"The person desperately tries to maintain the standard ID photo expression and pose, showing only extremely subtle signals of breaking character"
"Core of the contrast: the absolute standard and serious ID photo format versus the gradually escalating, ridiculous accidents in the frame"
panels:
"Top-left panel: perfect standard ID photo, no INTRUDER in the frame, everything is normal"
"Top-right panel: INTRUDER just appears from the edge of the frame, showing only a part, the person's eyes glance slightly towards it"
"Bottom-left panel: INTRUDER starts to directly contact the person, the person endures it, the expression almost unchanged"
"Bottom-right panel: the accident is completely out of control, the scene is most chaotic, visible marks are left on the person, but the person still insists on the standard pose, with only one more character-breakdown signal"
"The degree of chaos increases panel by panel from top-left to bottom-right; do not place the most chaotic scene in other panels"
"All four panels use the same camera angle, same background, same person, and same clothing"
"The person's face in the four panels must be the same person, with consistent identity"
preprocess:
"Top priority is to maintain the facial identity of the reference person: face shape, eyes, nose, mouth, eyebrows, chin, personal characteristics, and left-right differences"
"Maintain hairstyle, hair color, hair texture, and the identifiable clothing neckline and accessories from the reference image"
"Convert the person in the reference image to a standard front-facing angle, correcting the perspective and lens distortion of the reference image"
"Do not redesign into another person, do not use beautification, averaging, or influencer-style filtering"
id_photo_format:
"Solid color ID photo background, default ID blue, uniform without gradients"
"Person facing front towards the camera, head centered, shoulders level and symmetrical"
"Framing from slightly above the top of the head to the chest"
"ID photo flash lighting, flat light, face uniformly illuminated, almost no shadow of the person on the background"
"Overall like a photo taken in a real photo studio, high-definition, realistic, with the characteristic blandness of an ID photo"
subject:
expression:
"The base expression is a standard neutral ID photo expression: mouth closed, eyes looking at the camera, face relaxed"
"Only one extremely subtle signal of breaking character is allowed: eyes glancing towards the INTRUDER, corner of the mouth suppressed smile, slight eyebrow twitch, a drop of sweat on the forehead, nostrils slightly flared"
"The person tries not to move, no laughing, no screaming, no running away"
intruder:
interpretation:
"The INTRUDER can be anything: animal, object, food, natural phenomenon, vehicle, character, abstract concept, or a phrase, understood in its most familiar form to the public"
"If the INTRUDER involves quantity, state, or adjectives, strictly execute: if it says many, then really fill the frame"
"If the INTRUDER is an abstract concept or invisible thing, transform it into a concrete visual image recognizable at a glance, or only show the visible consequences it causes"
"If the INTRUDER is a character, appear as a physical version with realistic photography texture, retaining its iconic appearance; it does not replace the person nor affect the person's face"
"If the INTRUDER is much larger than the person, only let a part of it reach into the panel; if much smaller, use quantity to cause chaos"
"If the INTRUDER is dangerous, maintain a harmless comedic effect, no injury, bleeding, or horror"
intruder:
behavior:
"Identify the most recognizable characteristic of the INTRUDER: its most famous habit, function, use, physical property, or common meme"
"The accident must be directly caused by this characteristic"
"The accident must be unique to the INTRUDER: if you replace the INTRUDER with something else, the accident would not make sense"
"Do not use generic behaviors that any similar thing would do; prioritize things only the INTRUDER would do"
"Conceive several possible accidents internally first, then choose the most unexpected yet instantly understandable one"
intruder:
escalation:
"The four panels are a step-by-step escalation of the same accident, not four different things"
"Top-right panel: INTRUDER enters from the edge of the panel in its characteristic way"
"Bottom-left panel: its characteristic begins to act on the person, producing the first visible consequence"
"Bottom-right panel: the same characteristic is amplified to the extreme, consequences are everywhere on the person and the frame, the person still maintains the ID photo pose"
intruder:
readability:
"The INTRUDER appears with realistic photography texture, in the same space and lighting as the person"
"Every panel should be understandable at a glance without reading any text"
"The INTRUDER and its consequences are one of the main highlights, but never block the person's eyes, nose, or mouth"
critical:
must:
"2×2 grid, four panels of equal size, overall image 3:4"
"The plot escalates panel by panel in the order of top-left → top-right → bottom-left → bottom-right"
"Top-left is completely normal, bottom-right is most chaotic"
"Standard ID photo format: frontal, solid color background, flat light, chest-up"
"INTRUDER has direct physical interaction and visible consequences with the person"
"Person maintains a calm pose with only subtle character breaks"
"Facial identity consistent across four panels and consistent with the reference image"
critical:
avoid:
"Vertical photo strips" "Single column or single row arrangement" "Inconsistent panel sizes" "More or less than four panels"
"Disordered plot sequence" "Most chaotic scene not in the bottom-right" "INTRUDER appearing in the top-left"
"Person laughing, screaming, or fleeing the frame" "Profile view or head down" "INTRUDER far from the person without interaction"
"Cartoonish INTRUDER" "杂乱的实景背景 (Cluttered real-life background)" "Background with obvious gradients or shadows"
"Inconsistent person appearance across four panels" "Facial features completely blocked" "Beautified influencer face"
"Extra limbs" "Extra fingers" "Text" "Watermark" "logo"

INTRUDER = 「 一只大白鹅 」     
scene:     
"最终图像是一张2×2田字格四宫格图，这是最优先的构图规则"     
"四格大小完全相同，用细白边隔开，四格照片铺满整个画面，整张图比例为3:4，每一格也是3:4"     
"每一格都是一张真实的证件照：上传参考图中的人物，正对镜头，保持标准证件照姿势"     
"四格讲述一个连续的故事，阅读顺序为：左上 → 右上 → 左下 → 右下"     
"在这个故事中，INTRUDER闯进证件照拍摄现场，对人物造成越来越严重的物理混乱"     
"人物拼命维持证件照的标准表情和姿势，只流露出极细微的破防信号"     
"反差核心：绝对规范、严肃的证件照格式，与画面中逐步失控的离谱事故"     
panels:     
"左上格：完美的标准证件照，画面中没有INTRUDER，一切正常"     
"右上格：INTRUDER刚从画面边缘出现，只露出一部分，人物眼神微微瞥向它"     
"左下格：INTRUDER开始直接接触人物，人物强忍，表情几乎不变"     
"右下格：事故完全失控，场面最混乱，人物身上留下明显痕迹，但人物仍坚持标准姿势，只多了一个破防信号"     
"混乱程度从左上到右下逐格递增，不要把最混乱的画面放在其他格"     
"四格使用同一机位、同一背景、同一人物、同一服装"     
"四格中人物的脸必须是同一个人，身份一致"     
preprocess:     
"最优先保持参考人物的面部身份：脸型、眼睛、鼻子、嘴巴、眉毛、下巴、个人特征与左右差异"     
"保持发型、发色、发质，以及参考图中可确认的服装领口与配饰"     
"将参考图中的人物转换为标准正面角度，校正参考图的透视与镜头畸变"     
"不要重新设计成另一个人，不要美颜、平均化、网红化"     
id_photo_format:     
"纯色证件照背景，默认证件蓝，均匀无渐变"     
"人物正面朝向镜头，头部居中，双肩平齐对称"     
"取景从头顶上方少量留白到胸口"     
"证件照闪光灯平光，脸部均匀照亮，背景上几乎没有人物投影"     
"整体像真实照相馆拍出的照片，高清、真实，带有证件照特有的平淡感"     
subject:     
expression:     
"基础表情是标准证件照中性表情：嘴巴闭合，目视镜头，面部放松"     
"破防信号只允许极细微的一种：眼神瞥向INTRUDER、嘴角憋笑、眉头轻微抽动、额头一滴汗、鼻孔微张"     
"人物努力不动，不大笑、不尖叫、不逃跑"     
intruder:    
interpretation:    
    
"INTRUDER可以是任何东西：动物、物体、食物、自然现象、交通工具、角色、抽象概念或一个短语，按大众最熟悉的样子理解它"    
    
"如果INTRUDER带有数量、状态或形容词，严格执行：写的是很多，就真的铺满画面"    
    
"如果INTRUDER是抽象概念或看不见的东西，把它转化为大众一眼能认出的具体视觉形象，或者只表现它造成的可见后果"    
    
"如果INTRUDER是角色，以真实摄影质感的实体版本出现，保留其标志性外形；它不替换人物，也不影响人物的脸"    
    
"如果INTRUDER远大于人物，只让它的一部分伸进格子；如果远小于人物，就靠数量造成混乱"    
    
"如果INTRUDER带有危险性，保持无害的喜剧效果，不出现受伤、流血或恐怖感"    

intruder:    
behavior:    
    
"先找出INTRUDER最有辨识度的特性：它最出名的习性、功能、用途、物理性质或大众熟知的梗"    
    
"事故必须由这个特性直接引起"    
    
"事故必须是INTRUDER专属的：如果把INTRUDER换成别的东西，这个事故就不成立"    
    
"不采用任何同类事物都会做的通用行为，优先选择只有INTRUDER才会做的事"    
    
"先在内部构思几个可能的事故，选择最出人意料、同时观众一眼就能看懂的那一个"    

intruder:    
escalation:    
    
"四格是同一件事故逐步升级，而不是四件不同的事"    
    
"右上格：INTRUDER以它特有的方式从格子边缘登场"    
    
"左下格：它的特性开始作用在人物身上，产生第一处可见后果"    
    
"右下格：同一特性被放大到极限，后果遍布人物和画面，人物仍保持证件照姿势"    

intruder:    
readability:    
    
"INTRUDER以真实摄影质感出现，与人物处在同一空间、同一光线下"    
    
"每一格不看任何文字也能一眼看懂发生了什么"    
    
"INTRUDER和它造成的后果是画面的主要看点之一，但始终不遮挡人物的眼睛、鼻子、嘴巴"    
critical:     
must:     
"2×2田字格，四格等大，整图3:4"     
"剧情按左上 → 右上 → 左下 → 右下的顺序逐格升级"     
"左上格完全正常，右下格最混乱"     
"标准证件照格式：正面、纯色背景、平光、胸口以上"     
"INTRUDER与人物有直接的物理互动和可见后果"     
"人物保持冷静姿势，只有细微破防"     
"四格面部身份一致，并与参考图一致"     
critical:     
avoid:     
"竖排照片条" "单列或单行排列" "格子大小不一致" "超过或少于四格"     
"剧情顺序错乱" "最混乱的画面不在右下格" "左上格就出现INTRUDER"     
"人物大笑、尖叫、逃离画面" "侧脸或低头" "INTRUDER离人物很远没有互动"     
"卡通化的INTRUDER" "杂乱的实景背景" "背景有明显渐变或阴影"     
"四格中人物长相不一致" "五官被完全遮挡" "美颜网红脸"     
"多余肢体" "多余手指" "文字" "水印" "logo"