IMAGDressing
Interactive modular clothing generation
- Simple architecture: Generate realistic clothing and support user driven scene editing.
- Flexible plugin compatibility: Integrate with extension plugins such as IP Adapter, ControlNet, T2I Adapter, and AnimateDiff.
- Quick customization: Achieve quick customization in seconds without the need for additional LoRA training.
- Support text prompts: Support text prompts according to different scenarios.
- Support clothing replacement in designated areas (experimental function).
- Supports integration with ControlNet and IP Adapter.
- Provide IGPair dataset: containing over 300000 pairs of clothing and wearing images.
Product Details
IMAGDressing is an interactive modular clothing generation model designed to provide flexible and controllable customized services for virtual try on systems. This model integrates the semantic features of CLIP and the texture features of VAE into the denoised UNet using a hybrid attention module, ensuring that users can control editing. In addition, IMAGDressing also provides the IGPair dataset, which contains over 300000 pairs of clothing and wearing images, and establishes a standard data assembly process. This model can be combined with extension plugins such as ControlNet, IP Adapter, T2I Adapter, and AnimateDiff to enhance diversity and controllability.
