INF-LLaVA: Dual-perspective Perception for High-Resolution Multimodal Large Language Model

TL;DR

INF-LLaVA is a novel multimodal large language model that effectively processes high-resolution images by capturing both local details and global context through dual-perspective modules, outperforming existing models.

Contribution

The paper introduces INF-LLaVA, which employs dual-perspective cropping and enhancement modules to improve high-resolution image perception in multimodal models.

Findings

Outperforms existing MLLMs on multiple benchmarks

Effective high-resolution image processing with dual-perspective modules

Validated through extensive ablation studies

Abstract

With advancements in data availability and computing resources, Multimodal Large Language Models (MLLMs) have showcased capabilities across various fields. However, the quadratic complexity of the vision encoder in MLLMs constrains the resolution of input images. Most current approaches mitigate this issue by cropping high-resolution images into smaller sub-images, which are then processed independently by the vision encoder. Despite capturing sufficient local details, these sub-images lack global context and fail to interact with one another. To address this limitation, we propose a novel MLLM, INF-LLaVA, designed for effective high-resolution image perception. INF-LLaVA incorporates two innovative components. First, we introduce a Dual-perspective Cropping Module (DCM), which ensures that each sub-image contains continuous details from a local perspective and comprehensive information from a global perspective. Second, we introduce Dual-perspective Enhancement Module (DEM) to enable the mutual enhancement of global and local features, allowing INF-LLaVA to effectively process high-resolution images by simultaneously capturing detailed local information and comprehensive global context. Extensive ablation studies validate the effectiveness of these components, and experiments on a diverse set of benchmarks demonstrate that INF-LLaVA outperforms existing MLLMs. Code and pretrained model are available at https://github.com/WeihuangLin/INF-LLaVA.