SDDF：面向开放词表伪装目标检测的特异性驱动动态聚焦
SDDF: Specificity-Driven Dynamic Focusing for Open-Vocabulary Camouflaged Object Detection
针对伪装目标与背景视觉高度相似导致开放词表检测失效的问题，提出子描述主成分对比融合来过滤多模态大模型生成的冗余修饰词，并用特异性引导的区域弱对齐与动态聚焦增强前景-背景判别，可直接迁移到任何文本提示驱动的检测器中做提示去噪与区域对齐。
原文依据 we propose a specificity-guided regional weak alignment and dynamic focusing method
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论开放词表目标检测（OVOD）旨在借助文本提示检测开放世界中的已知与未知目标。受益于大规模视觉-语言预训练模型的出现，OVOD 已展现出较强的零样本泛化能力。然而，在处理伪装目标时，由于目标与背景的视觉特征高度相似，检测器往往无法区分并定位目标。为弥补这一空白，我们构建了一个名为 OVCOD-D 的基准，通过为精心挑选的伪装目标图像配以细粒度文本描述进行增强。由于现有伪装目标数据集规模有限，我们采用在大规模目标检测数据集上预训练的检测器作为基线方法，因为它们具备更强的零样本泛化能力。在多模态大模型生成的特异性感知子描述中，仍存在令人困惑和过度修饰的修饰词。为缓解此类干扰，我们设计了一种子描述主成分对比融合策略，以减少带噪的文本成分。此外，针对伪装目标视觉特征与其周围环境高度相似这一挑战，我们提出了一种特异性引导的区域弱对齐与动态聚焦方法，旨在增强检测器将伪装目标与背景区分开的能力。在开放集评估设置下，所提方法在 OVCOD-D 基准上取得 56.4 的 AP。项目代码与基准将在 https://github.com/Zh1fen/SDDF 发布。
英文原文摘要 1656 字符
Open-vocabulary object detection (OVOD) aims to detect known and unknown objects in the open world by leveraging text prompts. Benefiting from the emergence of large-scale vision--language pre-trained models, OVOD has demonstrated strong zero-shot generalization capabilities. However, when dealing with camouflaged objects, the detector often fails to distinguish and localize objects because the visual features of the objects and the background are highly similar. To bridge this gap, we construct a benchmark named OVCOD-D by augmenting carefully selected camouflaged object images with fine-grained textual descriptions. Due to the limited scale of available camouflaged object datasets, we adopt detectors pre-trained on large-scale object detection datasets as our baseline methods, as they possess stronger zero-shot generalization ability. In the specificity-aware sub-descriptions generated by multimodal large models, there still exist confusing and overly decorative modifiers. To mitigate such interference, we design a sub-description principal component contrastive fusion strategy that reduces noisy textual components. Furthermore, to address the challenge that the visual features of camouflaged objects are highly similar to those of their surrounding environment, we propose a specificity-guided regional weak alignment and dynamic focusing method, which aims to strengthen the detector's ability to discriminate camouflaged objects from background. Under the open-set evaluation setting, the proposed method achieves an AP of 56.4 on the OVCOD-D benchmark. Project code and benchmark will be released at https://github.com/Zh1fen/SDDF.