PixelFolder: An Efficient Progressive Pixel Synthesis Network for Image Generation

TL;DR

PixelFolder introduces a progressive pixel synthesis network that significantly reduces computational and parameter costs in image generation while achieving state-of-the-art results on benchmark datasets.

Contribution

It proposes a novel multi-stage pixel regression framework with pixel folding operations for efficient image synthesis, outperforming existing methods in speed and accuracy.

Findings

Reduces 89% computation compared to CIPS.

Achieves state-of-the-art FID scores on FFHQ and LSUN datasets.

More efficient than StyleGAN2 with 72% less computation.

Abstract

Pixel synthesis is a promising research paradigm for image generation, which can well exploit pixel-wise prior knowledge for generation. However, existing methods still suffer from excessive memory footprint and computation overhead. In this paper, we propose a progressive pixel synthesis network towards efficient image generation, coined as PixelFolder. Specifically, PixelFolder formulates image generation as a progressive pixel regression problem and synthesizes images via a multi-stage structure, which can greatly reduce the overhead caused by large tensor transformations. In addition, we introduce novel pixel folding operations to further improve model efficiency while maintaining pixel-wise prior knowledge for end-to-end regression. With these innovative designs, we greatly reduce the expenditure of pixel synthesis, e.g., reducing 89% computation and 53% parameters compared with the latest pixel synthesis method CIPS. To validate our approach, we conduct extensive experiments on two benchmark datasets, namely FFHQ and LSUN Church. The experimental results show that with much less expenditure, PixelFolder obtains new state-of-the-art (SOTA) performance on two benchmark datasets, i.e., 3.77 FID and 2.45 FID on FFHQ and LSUN Church, respectively.Meanwhile, PixelFolder is also more efficient than the SOTA methods like StyleGAN2, reducing about 72% computation and 31% parameters, respectively. These results greatly validate the effectiveness of the proposed PixelFolder.