Multimodal fashion image editing

Given a target model image, a sketch and a textual description of a garment, the aim is to generate a new image of the same model, wearing a new garment that follows the user input.