MMIU: Multimodal Multi-image Understanding for Evaluating Large Vision-Language Models
This addresses the problem of evaluating multi-image capabilities in LVLMs for researchers and developers, though it is incremental as it focuses on benchmarking rather than model innovation.
The authors tackled the lack of comprehensive evaluation for multi-image large vision-language models by introducing the MMIU benchmark, which includes 7 relationship types, 52 tasks, 77K images, and 11K questions, and found that even advanced models like GPT-4o achieve only 55.7% accuracy, revealing significant challenges in multi-image comprehension.
The capability to process multiple images is crucial for Large Vision-Language Models (LVLMs) to develop a more thorough and nuanced understanding of a scene. Recent multi-image LVLMs have begun to address this need. However, their evaluation has not kept pace with their development. To fill this gap, we introduce the Multimodal Multi-image Understanding (MMIU) benchmark, a comprehensive evaluation suite designed to assess LVLMs across a wide range of multi-image tasks. MMIU encompasses 7 types of multi-image relationships, 52 tasks, 77K images, and 11K meticulously curated multiple-choice questions, making it the most extensive benchmark of its kind. Our evaluation of 24 popular LVLMs, including both open-source and proprietary models, reveals significant challenges in multi-image comprehension, particularly in tasks involving spatial understanding. Even the most advanced models, such as GPT-4o, achieve only 55.7% accuracy on MMIU. Through multi-faceted analytical experiments, we identify key performance gaps and limitations, providing valuable insights for future model and data improvements. We aim for MMIU to advance the frontier of LVLM research and development, moving us toward achieving sophisticated multimodal multi-image user interactions.