Eliciting Complex Spatial Reasoning in MLLMs through Wide-Baseline Matching

Hugging Face Daily Papers Papers
spatial-reasoning multimodal-llms wide-baseline-matching benchmark reinforcement-learning reasoning

Summary

This paper introduces ReasonMatch-Bench, a benchmark for wide-baseline matching in multimodal LLMs, and proposes Dynamic Correspondence Reinforcement Learning (DCRL) to improve spatial reasoning. Experiments show significant gains on the benchmark while maintaining general performance.

Wide-baseline matching (WBM) requires integrating geometric understanding, viewpoint changes, fine-grained perception, and occlusion reasoning, making it a challenging testbed for spatial reasoning in multimodal large language models (MLLMs) deployed in physical environments. However, current MLLMs lack systematic evaluation and training frameworks for these capabilities. We introduce ReasonMatch-Bench, a benchmark stratified by viewpoint displacement and matching granularity across indoor, outdoor, and object-centric scenarios, and show that current MLLMs still struggle with fine-grained wide-baseline correspondence: on a difficult 90-sample subset, human annotators achieve 84.0 F1, while the best existing baseline reaches 37.2. To bridge this gap, we build a scalable data-generation pipeline that automatically extracts wide-baseline view pairs from large-scale video-3D corpora, including RGB-D videos and SfM reconstructions, yielding diverse and verifiable supervision. We further propose Dynamic Correspondence Reinforcement Learning (DCRL), which combines Image-Level Viewpoint Progression and Point-Level Correspondence Curriculum to improve WBM training through verifiable rewards without explicit CoT supervision. Extensive experiments show that DCRL substantially improves ReasonMatch-Bench and transfers to related spatial benchmarks, while maintaining general visual understanding performance with modest gains on several benchmarks.
Original Article
View Cached Full Text

Cached at: 06/04/26, 03:41 AM

Paper page - Eliciting Complex Spatial Reasoning in MLLMs through Wide-Baseline Matching

Source: https://huggingface.co/papers/2606.03577 Authors:

,

,

,

,

,

,

,

,

,

Abstract

Wide-baseline matching presents a challenging spatial reasoning testbed for multimodal large language models, requiring systematic evaluation and training frameworks that current models lack, prompting the introduction of ReasonMatch-Bench and Dynamic Correspondence Reinforcement Learning to improve performance.

Wide-baseline matching(WBM) requires integrating geometric understanding, viewpoint changes, fine-grained perception, and occlusion reasoning, making it a challenging testbed forspatial reasoninginmultimodal large language models(MLLMs) deployed in physical environments. However, current MLLMs lack systematic evaluation and training frameworks for these capabilities. We introduceReasonMatch-Bench, a benchmark stratified by viewpoint displacement and matching granularity across indoor, outdoor, and object-centric scenarios, and show that current MLLMs still struggle with fine-grained wide-baseline correspondence: on a difficult 90-sample subset, human annotators achieve 84.0 F1, while the best existing baseline reaches 37.2. To bridge this gap, we build a scalabledata-generation pipelinethat automatically extracts wide-baseline view pairs from large-scale video-3D corpora, includingRGB-D videosandSfM reconstructions, yielding diverse and verifiable supervision. We further proposeDynamic Correspondence Reinforcement Learning(DCRL), which combinesImage-Level Viewpoint ProgressionandPoint-Level Correspondence Curriculumto improve WBM training through verifiable rewards without explicit CoT supervision. Extensive experiments show that DCRL substantially improvesReasonMatch-Benchand transfers to related spatial benchmarks, while maintaining general visual understanding performance with modest gains on several benchmarks.

View arXiv pageView PDFProject pageGitHub7Add to collection

Get this paper in your agent:

hf papers read 2606\.03577

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.03577 in a model README.md to link it from this page.

Datasets citing this paper1

#### ReasonMatch/ReasonMatch Viewer• Updated6 minutes ago • 40

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.03577 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Reinforcing Dual-Path Reasoning in Spatial Vision Language Models

Hugging Face Daily Papers

This paper introduces SR-REAL, a unified framework for spatial vision-language models that combines linguistic deduction and 3D geometric reasoning via reinforcement learning, enabling robust multi-step spatial reasoning across diverse tasks.

Positional Failures in Long-Context LLMs: A Blind Spot in Reasoning Benchmarks

arXiv cs.CL

This paper identifies a blind spot in long-context LLM reasoning benchmarks: they fail to control task position within the context, allowing positional failures to go undetected. The authors propose Context Rot Evaluation (CRE) to systematically vary task position, filler content, and context length, revealing severe accuracy drops for some models when reasoning tasks are placed in the middle of long contexts.