arXiv cs.AI· Dan Zlotnikov, Ohad Ben-Shahar·· 12 小时前AI 评分24
DISTRACT：视觉退化下竞争目标间文本参考的退化指数研究
DISTRACT: Degradation-Indexed Study of Textual Reference Across Competing Targets
AI 导读
研究者提出 DISTRACT 基准，在固定视频、目标与干扰物的条件下，将四种文本先验与四种图像退化沿受控严重度轴交叉组合，评测五个视觉语言跟踪器（VLT）。Molmo2 与 JointNLT 表现出有方向的文本利用，UVLTrack、MMTrack、DUTrack 对文本语义变化响应微弱，且在 FDR 校正后无定向行为效应。
正文 · 原文
Abstract:Vision-language trackers (VLTs) accept descriptions intended to assist tracking, but accepting text does not establish that its content changes tracking decisions, particularly as visual evidence deteriorates. We introduce DISTRACT, a matched-intervention benchmark that holds video, target, and distractor fixed while crossing four text priors with four image degradations along a controlled severity axis. We evaluate five VLTs on four datasets using behavioral contrasts, layerwise representation probes, and activation patching. Molmo2 and JointNLT exhibit directed text use. Some effects contract under degradation, whereas others show no significant endpoint change after multiplicity correction. UVLTrack, MMTrack, and DUTrack respond only weakly to changes in text semantics, with no directed behavioral effect surviving FDR correction in the tested settings. Model-internal analyses reveal that at decision-near states, correct-versus-conflicting text is decodable in Molmo2 and JointNLT but not in the other three systems. In the responsive models, decodability peaks at late states where replacement produces little behavioral change, whereas replacing selected earlier states redirects conflict-associated trajectories toward the target. Because matched-correct donors do not significantly outperform unrelated donors, these interventions identify earlier intervention-sensitive states without establishing target-specific semantic rescue. DISTRACT thus exposes when and where text affects tracking under degraded visual evidence.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.28419 [cs.CV] |
| (or arXiv:2610.28419v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.28419 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Dan Zlotnikov [view email]
[v1]
Tue, 20 Oct 2026 15:53:23 UTC (8,758 KB)
来源：arXiv cs.AI · arxiv.org