提出新框架让模型在找不到伪装目标时主动放弃检测，避免误判。
CFCamo: A Counterfactual Detect-or-Abstain Framework for Camouflaged Object Detection
- 设计反事实配对训练机制，同时优化发现目标与放弃检测的能力。
- 在CAMO测试集上比基线提升3.7个百分点，反事实准确率达80%-90%。
- 适合需要高可信度检测的场景，如医疗影像或安防监控。
视觉-语言强化学习在伪装目标检测中表现出色，但现有方法仅关注目标存在时的定位能力。由于训练数据全为含目标图像，模型易产生过度检测偏差，即在无目标图像中仍误报目标存在，而传统评估无法衡量此行为。为此，我们构建了反事实伪装目标检测（CF-COD）基准，通过移除测试图像中的伪装目标并保留合理背景，形成原始图与反事实图配对。该基准以成对准确率（PA）衡量模型在原图检测目标、在反事实图正确放弃检测的能力。进一步提出CFCamo框架，采用反事实序列策略优化（CSPO）训练Qwen3-VL-4B-Instruct模型，利用反事实成对奖励（CPR）耦合原图检测与反事实弃权。在CAMO-test上，CFCamo相较先前基于强化学习的基线提升S_alpha达+3.7个百分点；在整体CF-COD上，实现80.0%-90.8%的成对准确率。消融实验表明，若去除反事实耦合，即使目标存在检测性能良好，成对准确率仍降至1.4%-5.2%，证明仅靠目标存在评估无法刻画检测-放弃行为。结果表明，CFCamo通过耦合目标存在检测与目标不存在弃权，真正提升了模型可靠性。代码与数据见：https://github.com/suhang2000/CFCamo。
原文摘要 · Abstract (English)
Vision-language reinforcement learning has recently shown strong target-present localization for camouflaged object detection (COD). Yet localization is only one side of the decision: when the agent faces an ordinary image with no camouflaged target, will it still claim that a camouflaged object exists? Standard COD training and evaluation data are positive-only, so agents optimized under this setting can acquire an over-detect bias, a task-specific form of object hallucination that standard COD evaluation leaves unmeasured. To quantify this target-absent behavior, we construct Counterfactual COD (CF-COD), a paired benchmark that removes the camouflaged target from each held-out COD evaluation image while preserving a plausible background. CF-COD evaluates whether a model detects the target on the original image and abstains on the target-absent counterfactual, summarized by Pair Accuracy (PA). We further introduce CFCamo, a paired counterfactual framework for COD with abstention. For training, CFCamo optimizes a Qwen3-VL-4B-Instruct agent with Counterfactual Sequence Policy Optimization (CSPO), which samples paired original-counterfactual rollouts and uses a Counterfactual Paired Reward (CPR) to couple original-image detection with counterfactual abstention. On CAMO-test, CFCamo improves S_alpha by +3.7 pp over the prior RL-based COD baseline; across CF-COD, it reaches 80.0-90.8% PA. Ablations show that removing counterfactual coupling reduces PA to 1.4-5.2% despite strong target-present COD scores, showing that target-present evaluation alone does not characterize detect-or-abstain behavior. Overall, these results indicate that CFCamo improves COD agents by coupling target-present detection with target-absent abstention, rather than merely strengthening target-present localization. Code and data are available at https://github.com/suhang2000/CFCamo.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。