Saved in:
Bibliographic Details
Main Authors: Li, Rang, Li, Lei, Ren, Shuhuai, Tian, Hao, Gu, Shuhao, Li, Shicheng, Yue, Zihao, Wang, Yudong, Ma, Wenhan, Yang, Zhe, Ma, Jingyuan, Sui, Zhifang, Luo, Fuli
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.17495
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910064509452288
author Li, Rang
Li, Lei
Ren, Shuhuai
Tian, Hao
Gu, Shuhao
Li, Shicheng
Yue, Zihao
Wang, Yudong
Ma, Wenhan
Yang, Zhe
Ma, Jingyuan
Sui, Zhifang
Luo, Fuli
author_facet Li, Rang
Li, Lei
Ren, Shuhuai
Tian, Hao
Gu, Shuhao
Li, Shicheng
Yue, Zihao
Wang, Yudong
Ma, Wenhan
Yang, Zhe
Ma, Jingyuan
Sui, Zhifang
Luo, Fuli
contents Visual grounding, localizing objects from natural language descriptions, represents a critical bridge between language and vision understanding. While multimodal large language models (MLLMs) achieve impressive scores on existing benchmarks, a fundamental question remains: can MLLMs truly visually ground with human-like sophistication, or are they merely pattern-matching on simplified datasets? Current benchmarks fail to capture real-world complexity where humans effortlessly navigate intricate references and recognize when grounding is impossible. To rigorously assess MLLMs' true capabilities, we introduce GroundingME, a benchmark that systematically challenges models across four critical dimensions: (1) Discriminative: distinguishing highly similar objects, (2) Spatial: understanding complex relational descriptions, (3) Limited: handling occlusions or tiny objects, and (4) Rejection: recognizing ungroundable queries. Through careful curation combining automated generation with human verification, we create 1,005 challenging examples mirroring real-world complexity. Evaluating 25 state-of-the-art MLLMs reveals a profound capability gap: the best model achieves only 45.1% accuracy, while most score 0% on rejection tasks. We explore two strategies for improvements: (1) test-time scaling selects optimal response by thinking trajectory to improve overall performance by up to 4.5%, and (2) data-mixture training boosts rejection accuracy from 0% to 27.9%. GroundingME thus serves as both a diagnostic tool revealing current limitations in MLLMs and a roadmap toward human-level visual grounding. Project page: https://groundingme.github.io
format Preprint
id arxiv_https___arxiv_org_abs_2512_17495
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GroundingME: Exposing the Visual Grounding Gap in MLLMs through Multi-Dimensional Evaluation
Li, Rang
Li, Lei
Ren, Shuhuai
Tian, Hao
Gu, Shuhao
Li, Shicheng
Yue, Zihao
Wang, Yudong
Ma, Wenhan
Yang, Zhe
Ma, Jingyuan
Sui, Zhifang
Luo, Fuli
Computer Vision and Pattern Recognition
Visual grounding, localizing objects from natural language descriptions, represents a critical bridge between language and vision understanding. While multimodal large language models (MLLMs) achieve impressive scores on existing benchmarks, a fundamental question remains: can MLLMs truly visually ground with human-like sophistication, or are they merely pattern-matching on simplified datasets? Current benchmarks fail to capture real-world complexity where humans effortlessly navigate intricate references and recognize when grounding is impossible. To rigorously assess MLLMs' true capabilities, we introduce GroundingME, a benchmark that systematically challenges models across four critical dimensions: (1) Discriminative: distinguishing highly similar objects, (2) Spatial: understanding complex relational descriptions, (3) Limited: handling occlusions or tiny objects, and (4) Rejection: recognizing ungroundable queries. Through careful curation combining automated generation with human verification, we create 1,005 challenging examples mirroring real-world complexity. Evaluating 25 state-of-the-art MLLMs reveals a profound capability gap: the best model achieves only 45.1% accuracy, while most score 0% on rejection tasks. We explore two strategies for improvements: (1) test-time scaling selects optimal response by thinking trajectory to improve overall performance by up to 4.5%, and (2) data-mixture training boosts rejection accuracy from 0% to 27.9%. GroundingME thus serves as both a diagnostic tool revealing current limitations in MLLMs and a roadmap toward human-level visual grounding. Project page: https://groundingme.github.io
title GroundingME: Exposing the Visual Grounding Gap in MLLMs through Multi-Dimensional Evaluation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.17495