计算机科学
水准点(测量)
隐藏字幕
人工智能
桥(图论)
帧(网络)
空格(标点符号)
空间分析
图像(数学)
编码(集合论)
机器学习
重写
动作(物理)
空间关系
代表(政治)
计算机视觉
空间智能
脆弱性(计算)
自然语言处理
空间语境意识
模式识别(心理学)
语言模型
语义学(计算机科学)
作者
Bozhou Li,Jiahang Zhang,Yue Ding,Yushuo Guan,Bohan Zeng,Yiyan Ji,Xinlong Chen,Yang Shi,Yifan Dai,Yuran Wang,Chengzhuo Tong,Pengfei Wan,Yuanxing Zhang,Wentao Zhang
出处
期刊:
[Association for Computing Machinery]
日期:2026-11-06
卷期号:: 4337-4346
标识
DOI:10.1145/3767308.3835525
摘要
While multimodal large language models (MLLMs) achieve remarkable performance on generic image captioning, they frequently suffer from structural hallucinations in human-centric scenarios. Accurately modeling human subjects is foundational for critical downstream applications, such as accurate avatar/video/image generation and fine-grained human action understanding. However, these tasks require highly precise subject-centered spatial grounding, such as distinguishing egocentric left/right laterality and maintaining correct anatomical-object bindings. Although catastrophic for structural integrity, these localized spatial inversions are often overshadowed by overall descriptive metrics in existing benchmarks. To systematically expose and quantify this bottleneck, we introduce SPACE (Subject-centric Poses, Appearance, and Characteristics Evaluation), a benchmark designed to evaluate subject-centered spatial understanding. On SPACE, we reveal that despite strong generic perception, current MLLMs consistently fail to ground descriptions in the subject's intrinsic frame of reference. To bridge this gap, we propose a specialized data construction and alignment pipeline. We first extract structured spatial hints from fine-grained body-part localization to guide a two-stage caption rewriting process, yielding highly spatially-faithful training data. Furthermore, we design a rubric-based reward for Group Relative Policy Optimization (GRPO) that explicitly penalizes structurally critical spatial errors during alignment. Extensive experiments on SPACE demonstrate our framework significantly improves human-centric caption quality, particularly in subject-centered spatial reasoning, achieving performance competitive with strong closed-source models. These results highlight the necessity of structured spatial alignment and set a new paradigm for generating fine-grained, anatomically grounded captions. Our benchmark and code are available at https://github.com/JHang2020/SPACE-Eval.
科研通智能强力驱动
Strongly Powered by AbleSci AI