Caption Injection for Optimization in Generative Search Engine

2025-11-06

AI Summary

Existing generative search engine optimization (G-SEO) methods are text-only and thus unable to leverage visual modalities—such as images—to enhance users’ subjective perceptibility of content. To address this limitation, we propose Caption Injection, the first multimodal G-SEO method tailored for generative search: it automatically extracts image captions and structurally injects them into textual content, enabling synergistic enhancement between visual semantics and language generation. Caption Injection is the first to integrate multimodal signals into the G-SEO pipeline, built upon a Multimodal Retrieval-Augmented Generation (MRAG) framework and rigorously evaluated on the MRAMG benchmark. Experimental results demonstrate statistically significant improvements over text-only baselines under the G-Eval metric, confirming that visual semantic injection effectively boosts user-perceived visibility of generated outputs. This work establishes a novel paradigm for multimodal search optimization.

Technology Category

• Computer Vision: Multi-modal Vision
• Search and Optimization: Learning to Search
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Generative Search Engines (GSEs) leverage Retrieval-Augmented Generation (RAG) techniques and Large Language Models (LLMs) to integrate multi-source information and provide users with accurate and comprehensive responses. Unlike traditional search engines that present results in ranked lists, GSEs shift users'attention from sequential browsing to content-driven subjective perception, driving a paradigm shift in information retrieval. In this context, enhancing the subjective visibility of content through Generative Search Engine Optimization (G-SEO) methods has emerged as a new research focus. With the rapid advancement of Multimodal Retrieval-Augmented Generation (MRAG) techniques, GSEs can now efficiently integrate text, images, audio, and video, producing richer responses that better satisfy complex information needs. Existing G-SEO methods, however, remain limited to text-based optimization and fail to fully exploit multimodal data. To address this gap, we propose Caption Injection, the first multimodal G-SEO approach, which extracts captions from images and injects them into textual content, integrating visual semantics to enhance the subjective visibility of content in generative search scenarios. We systematically evaluate Caption Injection on MRAMG, a benchmark for MRAG, under both unimodal and multimodal settings. Experimental results show that Caption Injection significantly outperforms text-only G-SEO baselines under the G-Eval metric, demonstrating the necessity and effectiveness of multimodal integration in G-SEO to improve user-perceived content visibility.

Problem

• Optimizing multimodal content visibility in generative search engines
• Integrating visual semantics through caption injection for G-SEO
• Addressing limitations of text-only optimization in multimodal retrieval systems

Innovation

• Proposes first multimodal G-SEO method called Caption Injection
• Extracts captions from images and injects into text
• Enhances content visibility in generative search scenarios