Workflow: Missing inputs
Missing required inputs: model try-on needs at least one garment reference and one model reference.
Input roles:
Garment reference count: 0
None
Model reference count: 0
None
Face identity reference count: 0
None
Scene reference count: 0
Pose action reference count: 0
Scene/composition reference count: 0
Style/atmosphere reference count: 0
Generation type: 模特穿着图
Create a commercial model try-on image. The garment reference controls the clothing. The model reference controls the final person/background base and human presentation. Dedicated pose references, when provided, control only human body action and joint relationships.
Execution rules:
Garment reference images determine clothing style, silhouette, color, pattern, fabric texture and local details. They have the highest priority.
Pose reference images are only for human body action and joint relationships. When present and enabled, they have priority over the original model-reference pose for body action only. They must not change subject identity, face, hairstyle, skin tone, body shape, garment style, background, props, scene, lighting, color tone, camera framing, composition or visual style.
Scene reference images are only for background, environment, spatial relationship, scene layout and framing. They must not change garment details, subject identity or styling.
Style reference images are only for color tone, lighting, atmosphere, mood and lens language. They must not change garment style or copy concrete background object structure.
Each task must use its explicitly paired garment reference and model reference. Do not randomly mix assets from other tasks.
Garment consistency has higher priority than scene creativity or beautification.
Do not change garment cut, neckline, sleeve, placket, buttons, hem, pattern content, pattern scale, pattern density or trim details.
Output must be realistic, clear and commercial-grade.
Apparel workflow fields:
Current mode: 精准换装 / 常规生成
Garment name: not provided
Garment category: 套装
Gender category: 女装
Age group: 成人
Special garment type: 睡衣
Reference priority:
The garment reference image has the highest priority and determines the garment style, color, structure, pattern, trims and material details.
The model reference image controls human identity, face, hairstyle, skin tone, body proportion, background, scene, camera, lighting and photographic realism. It may guide pose/body movement only when no dedicated pose reference is enabled. It must not override garment details.
The scene reference image only controls background, composition and lighting. It must not change the garment structure.
If any reference conflicts, preserve the garment reference first.
Pose generation mode: preserve or naturally refine the model reference pose only when no dedicated pose reference is enabled.
If a dedicated pose reference is enabled, transfer that pose reference's body action and joint relationships while keeping the model reference person and background.
A pose reference must not transfer its person identity, face, hairstyle, body shape, clothing, background, scene, lighting, color tone, camera framing, composition or visual style.
Sleepwear / homewear constraint: keep the garment soft, comfortable, lightweight and home-lounge oriented. Preserve pajama-set coordination and avoid turning it into outdoor fashion.
Matching set constraint: the top and bottom must come from the same garment set. Keep color, pattern, fabric and edge trims consistent across both pieces.