Skip to yearly menu bar Skip to main content


Poster

Compositional Text-to-Image Generation with Dense Blob Representations

Weili Nie · Sifei Liu · Morteza Mardani · Chao Liu · Benjamin Eckart · Arash Vahdat

Hall C 4-9 #410
[ ] [ Paper PDF ]
[ Poster
Tue 23 Jul 4:30 a.m. PDT — 6 a.m. PDT

Abstract:

Existing text-to-image models struggle to follow complex text prompts, raising the need for extra grounding inputs for better controllability. In this work, we propose to decompose a scene into visual primitives - denoted as dense blob representations - that contain fine-grained details of the scene while being modular, human-interpretable, and easy-to-construct. Based on blob representations, we develop a blob-grounded text-to-image diffusion model, termed BlobGEN, for compositional generation. Particularly, we introduce a new masked cross-attention module to disentangle the fusion between blob representations and visual features. To leverage the compositionality of large language models (LLMs), we introduce a new in-context learning approach to generate blob representations from text prompts. Our extensive experiments show that BlobGEN achieves superior zero-shot generation quality and better layout-guided controllability on MS-COCO. When augmented by LLMs, our method exhibits superior numerical and spatial correctness on compositional image generation benchmarks.

Chat is not available.