V*: Guided Visual Search as a Core Mechanism in Multimodal LLMs

TL;DR

This paper introduces V*, a visual search mechanism guided by language models, enhancing multimodal LLMs' ability to focus on important visual details in complex, high-resolution images, thereby improving reasoning and understanding.

Contribution

We propose V*, a novel LLM-guided visual search mechanism integrated into a new MLLM architecture called SEAL, addressing the lack of visual search in current models.

Findings

V* improves focus on visual details in high-resolution images.

SEAL architecture enhances reasoning and contextual understanding.

V*Bench evaluates MLLMs' visual search capabilities.

Abstract

When we look around and perform complex tasks, how we see and selectively process what we see is crucial. However, the lack of this visual search mechanism in current multimodal LLMs (MLLMs) hinders their ability to focus on important visual details, especially when handling high-resolution and visually crowded images. To address this, we introduce V*, an LLM-guided visual search mechanism that employs the world knowledge in LLMs for efficient visual querying. When combined with an MLLM, this mechanism enhances collaborative reasoning, contextual understanding, and precise targeting of specific visual elements. This integration results in a new MLLM meta-architecture, named Show, sEArch, and TelL (SEAL). We further create V*Bench, a benchmark specifically designed to evaluate MLLMs in their ability to process high-resolution images and focus on visual details. Our study highlights the necessity of incorporating visual search capabilities into multimodal systems. The code is available https://github.com/penghao-wu/vstar.