One of the most recent and striking innovations in machine learning is GAN (Generative Adversarial Network). GANs are generative models: they create new data instances that resemble your training data. For example, GANs can create images that look like photographs of human faces, even though the faces don’t belong to any actual person. GANs achieve this level of realism by pairing a generator, which learns to produce the target output, with a discriminator, which learns to distinguish accurate data from the outcome of the generator. The generator tries to fool the discriminator, and the discriminator tries to keep from being misled.
In Fashion, GAN is used to produce synthetic fashion items. Several types of research on using GAN to generate synthetic data with satisfying results [1]. Fashionic AI aims to use GAN to generate synthetic styles. With the help of Designer Style Classification, it can develop a new combination of clothes of a particular stylist or variety of stylists. It can also create its style as an AI Designer.
In November, we will transfer clothes on two different fashion models with Fashionic AI. We will reveal the performance of Fashionic AI with the images that our artificial intelligence engineers will publish. We will see more closely how the artificial intelligence module repairs the missing parts of the clothes during the change of clothes. In the near future, we will integrate the Fashionic AI module with the Inspirest Mobile APP.
Referances
[1] Zhang, H., Sun, Y., Liu, L., Wang, X., Li, L., & Liu, W. (2020). ClothingOut: a category-supervised GAN model for clothing segmentation and retrieval. Neural Computing and Applications, 32, 4519–4530. doi:10.1007/s00521–018–3691-y/
