DiffCollage: Parallel Generation of Large Content with Diffusion Models
Authors:
- 1 June 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 10188-10198
Abstract
We present DiffCollage, a compositional diffusion model that can generate large content by leveraging diffusion models trained on generating pieces of the large content. Our approach is based on a factor graph representation where each factor node represents a portion of the content and a variable node represents their overlap. This representation allows us to aggregate intermediate outputs from diffusion models defined on individual nodes to generate content of arbitrary size and shape in parallel without resorting to an autoregressive generation procedure. We apply DiffCollage to various tasks, including infinite image generation, panorama image generation, and long-duration text-guided motion generation. Extensive experimental results with a comparison to strong autoregressive baselines verify the effectiveness of our approach.Keywords
All Related Versions
This publication has 22 references indexed in Scilit:
- Aligning Latent and Image Spaces to Connect the Unconnectable2021
- Adversarial Generation of Continuous Images2021
- Image Generators with Conditionally-Independent Pixel Synthesis2021
- Taming Transformers for High-Resolution Image Synthesis2021
- COCO-GAN: Generation by Parts via Conditional Coordinating2019
- AMASS: Archive of Motion Capture As Surface Shapes2019
- Semantic Image Synthesis With Spatially-Adaptive Normalization2019
- A Connection Between Score Matching and Denoising AutoencodersNeural Computation, 2011
- ImageNet: A large-scale hierarchical image database2009
- Reverse-time diffusion equation modelsStochastic Processes and their Applications, 1982