← 研踪 PaperTrailarXiv:2611.19747 · v1
How Far Is Too Far? Object Recognition Declines Monotonically with Semantic Distance
首次提交 2026-11-15 11:21
版本更新 2026-11-15 11:21
雷达观测 2026-11-17 12:06
摘要
Vision models are sensitive to object-background context, yet it remains unclear how recognition changes as the semantic distance between objects and backgrounds increases. We introduce ImageNet-OOC1k, a benchmark enabling graded control of object-background relationships along a continuous consensus visual-semantic dissimilarity axis. This formulation reveals a highly consistent monotonic decline in recognition accuracy across 35 models spanning convolutional neural networks, Vision Transformers, and vision-language models. Internal analyses of Vision Transformers reveal patterns consistent with aggregation failure, where object evidence is present at the patch level but not consolidated into the final prediction. Using this characterization as a diagnostic probe, a simple spatial re-ranking strategy recovers a subset of these errors and improves in-sample accuracy by up to 5.7 percentage points. These results establish context sensitivity as a continuous and predictable property of modern vision systems, with Vision Transformer analyses suggesting that aggregation limitations may contribute to some failure types even when object evidence remains available.
作者备注:Accepted at NeurIPS 2026 (Spotlight)
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-15 11:21 · 最新观测版本
How Far Is Too Far? Object Recognition Declines Monotonically with Semantic Distance
Vision models are sensitive to object-background context, yet it remains unclear how recognition changes as the semantic distance between objects and backgrounds increases. We introduce ImageNet-OOC1k, a benchmark enabling graded control of object-background relationships along a continuous consensus visual-semantic dissimilarity axis. This formulation reveals a highly consistent monotonic decline in recognition accuracy across 35 models spanning convolutional neural networks, Vision Transformers, and vision-language models. Internal analyses of Vision Transformers reveal patterns consistent with aggregation failure, where object evidence is present at the patch level but not consolidated into the final prediction. Using this characterization as a diagnostic probe, a simple spatial re-ranking strategy recovers a subset of these errors and improves in-sample accuracy by up to 5.7 percentage points. These results establish context sensitivity as a continuous and predictable property of modern vision systems, with Vision Transformer analyses suggesting that aggregation limitations may contribute to some failure types even when object evidence remains available.
查看此版本解读 →查看此版本原文 ↗