Open Multimodal Retrieval-Augmented Factual Image Generation

TL;DR

This paper introduces ORIG, a retrieval-augmented framework for generating photorealistic images that are factually accurate, addressing limitations of existing models by iteratively retrieving and integrating web evidence.

Contribution

The paper proposes ORIG, a novel open multimodal retrieval-augmented approach for factual image generation, along with a new benchmark FIG-Eval for systematic evaluation.

Findings

ORIG significantly improves factual consistency in generated images.

ORIG outperforms strong baselines in image quality and factual accuracy.

The approach demonstrates potential for open knowledge integration in image synthesis.

Abstract

Large Multimodal Models (LMMs) have achieved remarkable progress in generating photorealistic and prompt-aligned images, but they often produce outputs that contradict verifiable knowledge, especially when prompts involve fine-grained attributes or time-sensitive events. Conventional retrieval-augmented approaches attempt to address this issue by introducing external information, yet they are fundamentally incapable of grounding generation in accurate and evolving knowledge due to their reliance on static sources and shallow evidence integration. To bridge this gap, we introduce ORIG, an agentic open multimodal retrieval-augmented framework for Factual Image Generation (FIG), a new task that requires both visual realism and factual grounding. ORIG iteratively retrieves and filters multimodal evidence from the web and incrementally integrates the refined knowledge into enriched prompts to guide generation. To support systematic evaluation, we build FIG-Eval, a benchmark spanning ten categories across perceptual, compositional, and temporal dimensions. Experiments demonstrate that ORIG substantially improves factual consistency and overall image quality over strong baselines, highlighting the potential of open multimodal retrieval for factual image generation.