Virtual try-on based on predetermined cloth patterns driven by generative ai

Abstract

Systems and methods for providing virtual try-on based on predetermined cloth patterns driven by generative AI are disclosed. An example method includes receiving an input image including an image of a body of a person and a first clothing worn by the person, receiving a text description of the first clothing and a further input image including a second clothing, encoding the input image into an image latent, generating a mask corresponding to the second clothing, encoding the text description into a text embedding, encoding the further input image into a clothing embedding, adding a noise to the image latent to obtain a noisy image latent, providing the noisy image latent, the text embedding, the clothing embedding, and the mask to a neural network to obtain a denoised image latent and decoding the denoised image latent into an output image including the body dressed in the second clothing.