All projects

Multimodal Text-to-Face Network

Generating human faces from natural-language descriptions with GANs.

May 2019 Generative AI

A generative system that synthesises human portraits directly from textual descriptions.

  • Built an automated system to generate human faces from textual descriptions using GANs.
  • Combined StackGAN + ProgressiveGAN architectures to generate 64×64 images.
  • Used InferSent with the GloVe model for text embeddings.
GANsStackGANProgressiveGANGloVe