ROVER: Routing Object-Centric Visual Evidence for Grounded Multi-Image Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lv, Guannan, Nie, Ren, Dou, Hongjian, Gao, Tingting
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911726025310208
author Lv, Guannan
Nie, Ren
Dou, Hongjian
Gao, Tingting
author_facet Lv, Guannan
Nie, Ren
Dou, Hongjian
Gao, Tingting
contents Multimodal Large Language Models (MLLMs) have increasingly localized and interleaved visual evidence for deliberative reasoning. Grounding-based approaches typically focus on regions of interest (RoIs) by injecting cropped image patches or RoI-specific features into the reasoning context. However, such designs can weaken holistic scene understanding and inter-object relations, while incurring decoding costs that scale with the number and size of RoIs. Alternatively, adaptive visual feature selection often requires fine-grained supervision or complex heuristics. To address these limitations, we propose ROVER (Routing Object-centric Visual Evidence for grounded multi-image Reasoning), a lightweight, learnable plugin for efficient global visual evidence routing. Upon each object grounding prediction, ROVER injects a step-specific token triplet to synergistically: (i) aggregate the ongoing reasoning context, (ii) distill intra-image cues into a visual working space via object-centric differential attention, and (iii) route and integrate history-aware evidence across objects and images within this space for subsequent reasoning. We integrate ROVER into Qwen2.5-VL-7B and develop an interleaved SFT-to-GRPO training pipeline. Strictly adhering to the original datasets and evaluation protocols, our method achieves the best performance on MM-GCoT (+4.8% answer accuracy, +14.6% grounding accuracy) and VideoEspresso (+8.6% answer accuracy). The VideoEspresso-trained model demonstrates strong transferability, outperforming the base model by +4.7% on average across diverse benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2605_27959
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ROVER: Routing Object-Centric Visual Evidence for Grounded Multi-Image Reasoning
Lv, Guannan
Nie, Ren
Dou, Hongjian
Gao, Tingting
Computer Vision and Pattern Recognition
Artificial Intelligence
Multimodal Large Language Models (MLLMs) have increasingly localized and interleaved visual evidence for deliberative reasoning. Grounding-based approaches typically focus on regions of interest (RoIs) by injecting cropped image patches or RoI-specific features into the reasoning context. However, such designs can weaken holistic scene understanding and inter-object relations, while incurring decoding costs that scale with the number and size of RoIs. Alternatively, adaptive visual feature selection often requires fine-grained supervision or complex heuristics. To address these limitations, we propose ROVER (Routing Object-centric Visual Evidence for grounded multi-image Reasoning), a lightweight, learnable plugin for efficient global visual evidence routing. Upon each object grounding prediction, ROVER injects a step-specific token triplet to synergistically: (i) aggregate the ongoing reasoning context, (ii) distill intra-image cues into a visual working space via object-centric differential attention, and (iii) route and integrate history-aware evidence across objects and images within this space for subsequent reasoning. We integrate ROVER into Qwen2.5-VL-7B and develop an interleaved SFT-to-GRPO training pipeline. Strictly adhering to the original datasets and evaluation protocols, our method achieves the best performance on MM-GCoT (+4.8% answer accuracy, +14.6% grounding accuracy) and VideoEspresso (+8.6% answer accuracy). The VideoEspresso-trained model demonstrates strong transferability, outperforming the base model by +4.7% on average across diverse benchmarks.
title ROVER: Routing Object-Centric Visual Evidence for Grounded Multi-Image Reasoning
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2605.27959