Enregistré dans:
Détails bibliographiques
Auteurs principaux: Zeng, Kang, Zhong, Guojin, Cheng, Jintao, Yuan, Jin, Li, Zhiyong
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2508.17860
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911120598499328
author Zeng, Kang
Zhong, Guojin
Cheng, Jintao
Yuan, Jin
Li, Zhiyong
author_facet Zeng, Kang
Zhong, Guojin
Cheng, Jintao
Yuan, Jin
Li, Zhiyong
contents The advancement of Multimodal Large Language Models (MLLMs) has driven significant progress in Visual Question Answering (VQA), evolving from Single to Multi Image VQA (MVQA). However, the increased number of images in MVQA inevitably introduces substantial visual redundancy that is irrelevant to question answering, negatively impacting both accuracy and efficiency. To address this issue, existing methods lack flexibility in controlling the number of compressed visual tokens and tend to produce discrete visual fragments, which hinder MLLMs' ability to comprehend images holistically. In this paper, we propose a straightforward yet universal Adaptive Visual Anchoring strategy, which can be seamlessly integrated into existing MLLMs, offering significant accuracy improvements through adaptive compression. Meanwhile, to balance the results derived from both global and compressed visual input, we further introduce a novel collaborative decoding mechanism, enabling optimal performance. Extensive experiments validate the effectiveness of our method, demonstrating consistent performance improvements across various MLLMs. The code will be publicly available.
format Preprint
id arxiv_https___arxiv_org_abs_2508_17860
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AVAM: Universal Training-free Adaptive Visual Anchoring Embedded into Multimodal Large Language Model for Multi-image Question Answering
Zeng, Kang
Zhong, Guojin
Cheng, Jintao
Yuan, Jin
Li, Zhiyong
Computer Vision and Pattern Recognition
Artificial Intelligence
The advancement of Multimodal Large Language Models (MLLMs) has driven significant progress in Visual Question Answering (VQA), evolving from Single to Multi Image VQA (MVQA). However, the increased number of images in MVQA inevitably introduces substantial visual redundancy that is irrelevant to question answering, negatively impacting both accuracy and efficiency. To address this issue, existing methods lack flexibility in controlling the number of compressed visual tokens and tend to produce discrete visual fragments, which hinder MLLMs' ability to comprehend images holistically. In this paper, we propose a straightforward yet universal Adaptive Visual Anchoring strategy, which can be seamlessly integrated into existing MLLMs, offering significant accuracy improvements through adaptive compression. Meanwhile, to balance the results derived from both global and compressed visual input, we further introduce a novel collaborative decoding mechanism, enabling optimal performance. Extensive experiments validate the effectiveness of our method, demonstrating consistent performance improvements across various MLLMs. The code will be publicly available.
title AVAM: Universal Training-free Adaptive Visual Anchoring Embedded into Multimodal Large Language Model for Multi-image Question Answering
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2508.17860