Timezone: »

Improving Visual Quality of Image Synthesis by A Token-based Generator with Transformers
Yanhong Zeng · Huan Yang · Hongyang Chao · Jianbo Wang · Jianlong Fu

Tue Dec 07 04:30 PM -- 06:00 PM (PST) @

We present a new perspective of achieving image synthesis by viewing this task as a visual token generation problem. Different from existing paradigms that directly synthesize a full image from a single input (e.g., a latent code), the new formulation enables a flexible local manipulation for different image regions, which makes it possible to learn content-aware and fine-grained style control for image synthesis. Specifically, it takes as input a sequence of latent tokens to predict the visual tokens for synthesizing an image. Under this perspective, we propose a token-based generator (i.e., TokenGAN). Particularly, the TokenGAN inputs two semantically different visual tokens, i.e., the learned constant content tokens and the style tokens from the latent space. Given a sequence of style tokens, the TokenGAN is able to control the image synthesis by assigning the styles to the content tokens by attention mechanism with a Transformer. We conduct extensive experiments and show that the proposed TokenGAN has achieved state-of-the-art results on several widely-used image synthesis benchmarks, including FFHQ and LSUN CHURCH with different resolutions. In particular, the generator is able to synthesize high-fidelity images with (1024x1024) size, dispensing with convolutions entirely.

Author Information

Yanhong Zeng (Sun Yat-sen University)
Huan Yang (Microsoft Research)
Hongyang Chao
Jianbo Wang (The University of Tokyo, Tokyo Institute of Technology)
Jianlong Fu (Microsoft Research)

More from the Same Authors