REVEALER: Reinforcement-Guided Visual Reasoning for Element-Level Text-Image Alignment Evaluation

2025-12-28

AI Summary

Text-to-image (T2I) generation lacks fine-grained, interpretable, and human-preference-aligned evaluation methods. Method: We propose a unified evaluation framework grounded in reinforcement-guided visual reasoning, introducing the novel “localize–reason–judge” paradigm to enable interpretable, element-level alignment quantification. Our approach integrates multimodal large language models (MLLMs) with visual grounding, designs a structured reward function, and introduces Group Relative Policy Optimization (GRPO)—a novel algorithm that jointly optimizes format compliance, localization accuracy, and alignment fidelity. Results: Our framework achieves state-of-the-art performance across four major benchmarks—EvalMuse-40K, RichHF, MHaluBench, and GenAI-Bench—significantly outperforming both proprietary models and supervised baselines. Moreover, it attains higher inference efficiency than iterative visual reasoning methods, enabling scalable, principled T2I evaluation.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Evaluating the alignment between textual prompts and generated images is critical for ensuring the reliability and usability of text-to-image (T2I) models. However, most existing evaluation methods rely on coarse-grained metrics or static QA pipelines, which lack fine-grained interpretability and struggle to reflect human preferences. To address this, we propose REVEALER, a unified framework for element-level alignment evaluation based on reinforcement-guided visual reasoning. Adopting a structured "grounding-reasoning-conclusion" paradigm, our method enables Multimodal Large Language Models (MLLMs) to explicitly localize semantic elements and derive interpretable alignment judgments. We optimize the model via Group Relative Policy Optimization(GRPO) using a composite reward function that incorporates structural format, grounding accuracy, and alignment fidelity. Extensive experiments across four benchmarks-EvalMuse-40K, RichHF, MHaluBench, and GenAI-Bench-demonstrate that REVEALER achieves state-of-the-art performance. Our approach consistently outperforms both strong proprietary models and supervised baselines while demonstrating superior inference efficiency compared to existing iterative visual reasoning methods.

Problem

• Evaluates element-level text-image alignment for T2I models
• Addresses coarse-grained metrics lacking fine-grained interpretability
• Proposes reinforcement-guided visual reasoning for interpretable judgments

Innovation

• Reinforcement-guided visual reasoning for fine-grained alignment
• Group Relative Policy Optimization with composite reward function
• Structured grounding-reasoning-conclusion paradigm using MLLMs