OutfitAnyone includes try-on model and auxiliary models. Use them together to meet a variety of business needs, from quick image generation and detail refinement to partial replacement.
This document applies only to the China (Beijing) region. To use the models, you must use an API key from the China (Beijing) region.
Model overview
Type | Service | Model | Core feature | Scenarios |
Try-on model | aitryon-plus | Improves image definition, clothing texture details, and logo restoration effects. | The generation takes a long time. This model is suitable for scenarios that are not time-sensitive. | |
Auxiliary model | aitryon-parsing-v1 | Segments the clothing area in an image. | Used to implement partial try-on or obtain clothing coordinates for interactions such as product hotspots. |
Scenarios
You can combine and call these APIs based on your business needs. The following sections describe the call flows for common scenarios:
Scenario 1: Basic try-on
Scenarios: For quickly verifying results or for business needs that require high-speed image generation.
Steps:
Prepare inputs: Prepare a model image and a clothing image.
Call try-on model: Call OutfitAnyone-Plus, and pass the two images as the core parameters.
Obtain the result: The API directly returns the try-on image.
Scenario 2: Partial try-on
Scenarios: Replacing a part of the outfit. For example, keeping the model's original pants and replacing only the top.
Steps:
Extract the clothing to keep: Call OutfitAnyone-Image Parsing, input the original model image, and obtain an image of the clothing item to keep (for example, the pants).
Combine for try-on: Call OutfitAnyone-Plus, and pass the original model image, the new top image, and the pants image obtained in the previous step.
Obtain the result: The API returns a try-on image that combines the new top and the original pants.
Scenario 3: Obtaining clothing coordinates
Scenarios: Adding product labels to the generated image or implementing interactive features such as product hotspots.
Steps:
Call the image segmentation model: Call OutfitAnyone-Image Parsing and input any model image (you can use a real photo or an image generated by OutfitAnyone).
Obtain the result: The API returns the coordinates (bbox) of the clothing area and a visualization. You can use these coordinates to implement frontend interactive features, such as product hotspots.