UNIMO-G: Unified Image Generation through Multimodal Conditional Diffusion

TL;DR

UNIMO-G introduces a unified multimodal diffusion framework that effectively generates high-fidelity images from complex prompts combining text and visual inputs, advancing both text-driven and subject-driven image synthesis.

Contribution

It presents a novel multimodal conditional diffusion model with a two-stage training process for unified image generation from multimodal prompts.

Findings

Excels in text-to-image and zero-shot subject-driven synthesis.

Generates high-fidelity images from complex multimodal prompts.

Effective in handling multiple image entities in prompts.

Abstract

Existing text-to-image diffusion models primarily generate images from text prompts. However, the inherent conciseness of textual descriptions poses challenges in faithfully synthesizing images with intricate details, such as specific entities or scenes. This paper presents UNIMO-G, a simple multimodal conditional diffusion framework that operates on multimodal prompts with interleaved textual and visual inputs, which demonstrates a unified ability for both text-driven and subject-driven image generation. UNIMO-G comprises two core components: a Multimodal Large Language Model (MLLM) for encoding multimodal prompts, and a conditional denoising diffusion network for generating images based on the encoded multimodal input. We leverage a two-stage training strategy to effectively train the framework: firstly pre-training on large-scale text-image pairs to develop conditional image generation capabilities, and then instruction tuning with multimodal prompts to achieve unified image generation proficiency. A well-designed data processing pipeline involving language grounding and image segmentation is employed to construct multi-modal prompts. UNIMO-G excels in both text-to-image generation and zero-shot subject-driven synthesis, and is notably effective in generating high-fidelity images from complex multimodal prompts involving multiple image entities.